generic-library/opencv
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Merge remote-tracking branch 'upstream/master' into HomographyDecomp

Browse Source
This commit is contained in:
Samson Yilma
2014-07-30 20:36:41 -04:00
parent 3b608fa489 1a1097ab23
commit 4fe04775d1
971 changed files with 74086 additions and 164430 deletions
Download Patch File Download Diff File Expand all files Collapse all files

32
modules/calib3d/doc/camera_calibration_and_3d_reconstruction.rst
View File

@@ -224,9 +224,9 @@ Computes useful camera characteristics from the camera matrix.
:param imageSize: Input image size in pixels.
:param apertureWidth: Physical width of the sensor.
:param apertureWidth: Physical width in mm of the sensor.
:param apertureHeight: Physical height of the sensor.
:param apertureHeight: Physical height in mm of the sensor.
:param fovx: Output field of view in degrees along the horizontal sensor axis.
@@ -234,13 +234,15 @@ Computes useful camera characteristics from the camera matrix.
:param focalLength: Focal length of the lens in mm.
:param principalPoint: Principal point in pixels.
:param principalPoint: Principal point in mm.
:param aspectRatio: :math:`f_y/f_x`
The function computes various useful camera characteristics from the previously estimated camera matrix.
.. note::
Do keep in mind that the unity measure 'mm' stands for whatever unit of measure one chooses for the chessboard pitch (it can thus be any value).
composeRT
-------------
@@ -582,15 +584,15 @@ Finds an object pose from 3D-2D point correspondences.
:param flags: Method for solving a PnP problem:
* **CV_ITERATIVE** Iterative method is based on Levenberg-Marquardt optimization. In this case the function finds such a pose that minimizes reprojection error, that is the sum of squared distances between the observed projections ``imagePoints`` and the projected (using :ocv:func:`projectPoints` ) ``objectPoints`` .
* **CV_P3P** Method is based on the paper of X.S. Gao, X.-R. Hou, J. Tang, H.-F. Chang "Complete Solution Classification for the Perspective-Three-Point Problem". In this case the function requires exactly four object and image points.
* **CV_EPNP** Method has been introduced by F.Moreno-Noguer, V.Lepetit and P.Fua in the paper "EPnP: Efficient Perspective-n-Point Camera Pose Estimation".
* **ITERATIVE** Iterative method is based on Levenberg-Marquardt optimization. In this case the function finds such a pose that minimizes reprojection error, that is the sum of squared distances between the observed projections ``imagePoints`` and the projected (using :ocv:func:`projectPoints` ) ``objectPoints`` .
* **P3P** Method is based on the paper of X.S. Gao, X.-R. Hou, J. Tang, H.-F. Chang "Complete Solution Classification for the Perspective-Three-Point Problem". In this case the function requires exactly four object and image points.
* **EPNP** Method has been introduced by F.Moreno-Noguer, V.Lepetit and P.Fua in the paper "EPnP: Efficient Perspective-n-Point Camera Pose Estimation".
The function estimates the object pose given a set of object points, their corresponding image projections, as well as the camera matrix and the distortion coefficients.
.. note::
* An example of how to use solvePNP for planar augmented reality can be found at opencv_source_code/samples/python2/plane_ar.py
* An example of how to use solvePnP for planar augmented reality can be found at opencv_source_code/samples/python2/plane_ar.py
solvePnPRansac
------------------
@@ -707,8 +709,8 @@ Calculates an essential matrix from the corresponding points in two images.
:param method: Method for computing a fundamental matrix.
* **CV_RANSAC** for the RANSAC algorithm.
* **CV_LMEDS** for the LMedS algorithm.
* **RANSAC** for the RANSAC algorithm.
* **MEDS** for the LMedS algorithm.
:param threshold: Parameter used for RANSAC. It is the maximum distance from a point to an epipolar line in pixels, beyond which the point is considered an outlier and is not used for computing the final fundamental matrix. It can be set to something like 1-3, depending on the accuracy of the point localization, image resolution, and the image noise.
@@ -828,7 +830,7 @@ In this scenario, ``points1`` and ``points2`` are the same input for ``findEssen
cv::Point2d pp(0.0, 0.0);
Mat E, R, t, mask;
E = findEssentialMat(points1, points2, focal, pp, CV_RANSAC, 0.999, 1.0, mask);
E = findEssentialMat(points1, points2, focal, pp, RANSAC, 0.999, 1.0, mask);
recoverPose(E, points1, points2, R, t, focal, pp, mask);
@@ -851,9 +853,9 @@ Finds a perspective transformation between two planes.
* **0** - a regular method using all the points
* **CV_RANSAC** - RANSAC-based robust method
* **RANSAC** - RANSAC-based robust method
* **CV_LMEDS** - Least-Median robust method
* **LMEDS** - Least-Median robust method
:param ransacReprojThreshold: Maximum allowed reprojection error to treat a point pair as an inlier (used in the RANSAC method only). That is, if
@@ -863,7 +865,7 @@ Finds a perspective transformation between two planes.
then the point :math:`i` is considered an outlier. If ``srcPoints`` and ``dstPoints`` are measured in pixels, it usually makes sense to set this parameter somewhere in the range of 1 to 10.
:param mask: Optional output mask set by a robust method ( ``CV_RANSAC`` or ``CV_LMEDS`` ). Note that the input mask values are ignored.
:param mask: Optional output mask set by a robust method ( ``RANSAC`` or ``LMEDS`` ). Note that the input mask values are ignored.
The functions find and return the perspective transformation :math:`H` between the source and the destination planes:
@@ -1511,6 +1513,10 @@ Reconstructs points by triangulation.
The function reconstructs 3-dimensional points (in homogeneous coordinates) by using their observations with a stereo camera. Projections matrices can be obtained from :ocv:func:`stereoRectify`.
.. note::
Keep in mind that all input data should be of float type in order for this function to work.
.. seealso::
:ocv:func:`reprojectImageTo3D`

2
modules/calib3d/perf/perf_precomp.hpp
View File

@@ -11,7 +11,7 @@
#include "opencv2/ts.hpp"
#include "opencv2/calib3d.hpp"
#include "opencv2/highgui.hpp"
#include "opencv2/imgcodecs.hpp"
#include "opencv2/imgproc.hpp"
#ifdef GTEST_CREATE_SHARED_LIBRARY

2
modules/calib3d/src/solvepnp.cpp
View File

@@ -94,7 +94,7 @@ bool cv::solvePnP( InputArray _opoints, InputArray _ipoints,
return true;
}
else
CV_Error(CV_StsBadArg, "The flags argument must be one of CV_ITERATIVE or CV_EPNP");
CV_Error(CV_StsBadArg, "The flags argument must be one of CV_ITERATIVE, CV_P3P or CV_EPNP");
return false;
}

26
modules/calib3d/src/stereosgbm.cpp
View File

@@ -1029,18 +1029,6 @@ void filterSpecklesImpl(cv::Mat& img, int newVal, int maxSpeckleSize, int maxDif
T dp = *dpp;
int* lpp = labels + width*p.y + p.x;
if( p.x < width-1 && !lpp[+1] && dpp[+1] != newVal && std::abs(dp - dpp[+1]) <= maxDiff )
{
lpp[+1] = curlabel;
*ws++ = Point2s(p.x+1, p.y);
}
if( p.x > 0 && !lpp[-1] && dpp[-1] != newVal && std::abs(dp - dpp[-1]) <= maxDiff )
{
lpp[-1] = curlabel;
*ws++ = Point2s(p.x-1, p.y);
}
if( p.y < height-1 && !lpp[+width] && dpp[+dstep] != newVal && std::abs(dp - dpp[+dstep]) <= maxDiff )
{
lpp[+width] = curlabel;
@@ -1053,6 +1041,18 @@ void filterSpecklesImpl(cv::Mat& img, int newVal, int maxSpeckleSize, int maxDif
*ws++ = Point2s(p.x, p.y-1);
}
if( p.x < width-1 && !lpp[+1] && dpp[+1] != newVal && std::abs(dp - dpp[+1]) <= maxDiff )
{
lpp[+1] = curlabel;
*ws++ = Point2s(p.x+1, p.y);
}
if( p.x > 0 && !lpp[-1] && dpp[-1] != newVal && std::abs(dp - dpp[-1]) <= maxDiff )
{
lpp[-1] = curlabel;
*ws++ = Point2s(p.x-1, p.y);
}
// pop most recent and propagate
// NB: could try least recent, maybe better convergence
p = *--ws;
@@ -1084,7 +1084,7 @@ void cv::filterSpeckles( InputOutputArray _img, double _newval, int maxSpeckleSi
int newVal = cvRound(_newval), maxDiff = cvRound(_maxDiff);
#if defined HAVE_IPP && !defined HAVE_IPP_ICV_ONLY
#if IPP_VERSION_X100 >= 801
Ipp32s bufsize = 0;
IppiSize roisize = { img.cols, img.rows };
IppDataType datatype = type == CV_8UC1 ? ipp8u : ipp16s;

2
modules/calib3d/test/test_precomp.hpp
View File

@@ -13,7 +13,7 @@
#include "opencv2/ts.hpp"
#include "opencv2/imgproc.hpp"
#include "opencv2/calib3d.hpp"
#include "opencv2/highgui.hpp"
#include "opencv2/imgcodecs.hpp"
namespace cvtest
{

1
modules/contrib/CMakeLists.txt
View File

@@ -1 +0,0 @@
ocv_define_module(contrib opencv_imgproc opencv_calib3d opencv_ml opencv_video opencv_objdetect OPTIONAL opencv_highgui opencv_nonfree)

12
modules/contrib/doc/contrib.rst
View File

@@ -1,12 +0,0 @@
***************************************
contrib. Contributed/Experimental Stuff
***************************************
The module contains some recently added functionality that has not been stabilized, or functionality that is considered optional.
.. toctree::
:maxdepth: 2
stereo
FaceRecognizer Documentation <facerec/index>
openfabmap

107
modules/contrib/doc/facerec/colormaps.rst
View File

@@ -1,107 +0,0 @@
ColorMaps in OpenCV
===================
applyColorMap
---------------------
Applies a GNU Octave/MATLAB equivalent colormap on a given image.
.. ocv:function:: void applyColorMap(InputArray src, OutputArray dst, int colormap)
:param src: The source image, grayscale or colored does not matter.
:param dst: The result is the colormapped source image. Note: :ocv:func:`Mat::create` is called on dst.
:param colormap: The colormap to apply, see the list of available colormaps below.
Currently the following GNU Octave/MATLAB equivalent colormaps are implemented:
.. code-block:: cpp
enum
{
COLORMAP_AUTUMN = 0,
COLORMAP_BONE = 1,
COLORMAP_JET = 2,
COLORMAP_WINTER = 3,
COLORMAP_RAINBOW = 4,
COLORMAP_OCEAN = 5,
COLORMAP_SUMMER = 6,
COLORMAP_SPRING = 7,
COLORMAP_COOL = 8,
COLORMAP_HSV = 9,
COLORMAP_PINK = 10,
COLORMAP_HOT = 11
}
Description
-----------
The human perception isn't built for observing fine changes in grayscale images. Human eyes are more sensitive to observing changes between colors, so you often need to recolor your grayscale images to get a clue about them. OpenCV now comes with various colormaps to enhance the visualization in your computer vision application.
In OpenCV 2.4 you only need :ocv:func:`applyColorMap` to apply a colormap on a given image. The following sample code reads the path to an image from command line, applies a Jet colormap on it and shows the result:
.. code-block:: cpp
#include <opencv2/contrib.hpp>
#include <opencv2/core.hpp>
#include <opencv2/highgui.hpp>
using namespace cv;
int main(int argc, const char *argv[]) {
// Get the path to the image, if it was given
// if no arguments were given.
String filename;
if (argc > 1) {
filename = String(argv[1]);
}
// The following lines show how to apply a colormap on a given image
// and show it with cv::imshow example with an image. An exception is
// thrown if the path to the image is invalid.
if(!filename.empty()) {
Mat img0 = imread(filename);
// Throw an exception, if the image can't be read:
if(img0.empty()) {
CV_Error(CV_StsBadArg, "Sample image is empty. Please adjust your path, so it points to a valid input image!");
}
// Holds the colormap version of the image:
Mat cm_img0;
// Apply the colormap:
applyColorMap(img0, cm_img0, COLORMAP_JET);
// Show the result:
imshow("cm_img0", cm_img0);
waitKey(0);
}
return 0;
}
And here are the color scales for each of the available colormaps:
+-----------------------+---------------------------------------------------+
| Class | Scale |
+=======================+===================================================+
| COLORMAP_AUTUMN | .. image:: img/colormaps/colorscale_autumn.jpg |
+-----------------------+---------------------------------------------------+
| COLORMAP_BONE | .. image:: img/colormaps/colorscale_bone.jpg |
+-----------------------+---------------------------------------------------+
| COLORMAP_COOL | .. image:: img/colormaps/colorscale_cool.jpg |
+-----------------------+---------------------------------------------------+
| COLORMAP_HOT | .. image:: img/colormaps/colorscale_hot.jpg |
+-----------------------+---------------------------------------------------+
| COLORMAP_HSV | .. image:: img/colormaps/colorscale_hsv.jpg |
+-----------------------+---------------------------------------------------+
| COLORMAP_JET | .. image:: img/colormaps/colorscale_jet.jpg |
+-----------------------+---------------------------------------------------+
| COLORMAP_OCEAN | .. image:: img/colormaps/colorscale_ocean.jpg |
+-----------------------+---------------------------------------------------+
| COLORMAP_PINK | .. image:: img/colormaps/colorscale_pink.jpg |
+-----------------------+---------------------------------------------------+
| COLORMAP_RAINBOW | .. image:: img/colormaps/colorscale_rainbow.jpg |
+-----------------------+---------------------------------------------------+
| COLORMAP_SPRING | .. image:: img/colormaps/colorscale_spring.jpg |
+-----------------------+---------------------------------------------------+
| COLORMAP_SUMMER | .. image:: img/colormaps/colorscale_summer.jpg |
+-----------------------+---------------------------------------------------+
| COLORMAP_WINTER | .. image:: img/colormaps/colorscale_winter.jpg |
+-----------------------+---------------------------------------------------+

400
modules/contrib/doc/facerec/etc/at.txt
View File

@@ -1,400 +0,0 @@
/home/philipp/facerec/data/at/s13/2.pgm;12
/home/philipp/facerec/data/at/s13/7.pgm;12
/home/philipp/facerec/data/at/s13/6.pgm;12
/home/philipp/facerec/data/at/s13/9.pgm;12
/home/philipp/facerec/data/at/s13/5.pgm;12
/home/philipp/facerec/data/at/s13/3.pgm;12
/home/philipp/facerec/data/at/s13/4.pgm;12
/home/philipp/facerec/data/at/s13/10.pgm;12
/home/philipp/facerec/data/at/s13/8.pgm;12
/home/philipp/facerec/data/at/s13/1.pgm;12
/home/philipp/facerec/data/at/s17/2.pgm;16
/home/philipp/facerec/data/at/s17/7.pgm;16
/home/philipp/facerec/data/at/s17/6.pgm;16
/home/philipp/facerec/data/at/s17/9.pgm;16
/home/philipp/facerec/data/at/s17/5.pgm;16
/home/philipp/facerec/data/at/s17/3.pgm;16
/home/philipp/facerec/data/at/s17/4.pgm;16
/home/philipp/facerec/data/at/s17/10.pgm;16
/home/philipp/facerec/data/at/s17/8.pgm;16
/home/philipp/facerec/data/at/s17/1.pgm;16
/home/philipp/facerec/data/at/s32/2.pgm;31
/home/philipp/facerec/data/at/s32/7.pgm;31
/home/philipp/facerec/data/at/s32/6.pgm;31
/home/philipp/facerec/data/at/s32/9.pgm;31
/home/philipp/facerec/data/at/s32/5.pgm;31
/home/philipp/facerec/data/at/s32/3.pgm;31
/home/philipp/facerec/data/at/s32/4.pgm;31
/home/philipp/facerec/data/at/s32/10.pgm;31
/home/philipp/facerec/data/at/s32/8.pgm;31
/home/philipp/facerec/data/at/s32/1.pgm;31
/home/philipp/facerec/data/at/s10/2.pgm;9
/home/philipp/facerec/data/at/s10/7.pgm;9
/home/philipp/facerec/data/at/s10/6.pgm;9
/home/philipp/facerec/data/at/s10/9.pgm;9
/home/philipp/facerec/data/at/s10/5.pgm;9
/home/philipp/facerec/data/at/s10/3.pgm;9
/home/philipp/facerec/data/at/s10/4.pgm;9
/home/philipp/facerec/data/at/s10/10.pgm;9
/home/philipp/facerec/data/at/s10/8.pgm;9
/home/philipp/facerec/data/at/s10/1.pgm;9
/home/philipp/facerec/data/at/s27/2.pgm;26
/home/philipp/facerec/data/at/s27/7.pgm;26
/home/philipp/facerec/data/at/s27/6.pgm;26
/home/philipp/facerec/data/at/s27/9.pgm;26
/home/philipp/facerec/data/at/s27/5.pgm;26
/home/philipp/facerec/data/at/s27/3.pgm;26
/home/philipp/facerec/data/at/s27/4.pgm;26
/home/philipp/facerec/data/at/s27/10.pgm;26
/home/philipp/facerec/data/at/s27/8.pgm;26
/home/philipp/facerec/data/at/s27/1.pgm;26
/home/philipp/facerec/data/at/s5/2.pgm;4
/home/philipp/facerec/data/at/s5/7.pgm;4
/home/philipp/facerec/data/at/s5/6.pgm;4
/home/philipp/facerec/data/at/s5/9.pgm;4
/home/philipp/facerec/data/at/s5/5.pgm;4
/home/philipp/facerec/data/at/s5/3.pgm;4
/home/philipp/facerec/data/at/s5/4.pgm;4
/home/philipp/facerec/data/at/s5/10.pgm;4
/home/philipp/facerec/data/at/s5/8.pgm;4
/home/philipp/facerec/data/at/s5/1.pgm;4
/home/philipp/facerec/data/at/s20/2.pgm;19
/home/philipp/facerec/data/at/s20/7.pgm;19
/home/philipp/facerec/data/at/s20/6.pgm;19
/home/philipp/facerec/data/at/s20/9.pgm;19
/home/philipp/facerec/data/at/s20/5.pgm;19
/home/philipp/facerec/data/at/s20/3.pgm;19
/home/philipp/facerec/data/at/s20/4.pgm;19
/home/philipp/facerec/data/at/s20/10.pgm;19
/home/philipp/facerec/data/at/s20/8.pgm;19
/home/philipp/facerec/data/at/s20/1.pgm;19
/home/philipp/facerec/data/at/s30/2.pgm;29
/home/philipp/facerec/data/at/s30/7.pgm;29
/home/philipp/facerec/data/at/s30/6.pgm;29
/home/philipp/facerec/data/at/s30/9.pgm;29
/home/philipp/facerec/data/at/s30/5.pgm;29
/home/philipp/facerec/data/at/s30/3.pgm;29
/home/philipp/facerec/data/at/s30/4.pgm;29
/home/philipp/facerec/data/at/s30/10.pgm;29
/home/philipp/facerec/data/at/s30/8.pgm;29
/home/philipp/facerec/data/at/s30/1.pgm;29
/home/philipp/facerec/data/at/s39/2.pgm;38
/home/philipp/facerec/data/at/s39/7.pgm;38
/home/philipp/facerec/data/at/s39/6.pgm;38
/home/philipp/facerec/data/at/s39/9.pgm;38
/home/philipp/facerec/data/at/s39/5.pgm;38
/home/philipp/facerec/data/at/s39/3.pgm;38
/home/philipp/facerec/data/at/s39/4.pgm;38
/home/philipp/facerec/data/at/s39/10.pgm;38
/home/philipp/facerec/data/at/s39/8.pgm;38
/home/philipp/facerec/data/at/s39/1.pgm;38
/home/philipp/facerec/data/at/s35/2.pgm;34
/home/philipp/facerec/data/at/s35/7.pgm;34
/home/philipp/facerec/data/at/s35/6.pgm;34
/home/philipp/facerec/data/at/s35/9.pgm;34
/home/philipp/facerec/data/at/s35/5.pgm;34
/home/philipp/facerec/data/at/s35/3.pgm;34
/home/philipp/facerec/data/at/s35/4.pgm;34
/home/philipp/facerec/data/at/s35/10.pgm;34
/home/philipp/facerec/data/at/s35/8.pgm;34
/home/philipp/facerec/data/at/s35/1.pgm;34
/home/philipp/facerec/data/at/s23/2.pgm;22
/home/philipp/facerec/data/at/s23/7.pgm;22
/home/philipp/facerec/data/at/s23/6.pgm;22
/home/philipp/facerec/data/at/s23/9.pgm;22
/home/philipp/facerec/data/at/s23/5.pgm;22
/home/philipp/facerec/data/at/s23/3.pgm;22
/home/philipp/facerec/data/at/s23/4.pgm;22
/home/philipp/facerec/data/at/s23/10.pgm;22
/home/philipp/facerec/data/at/s23/8.pgm;22
/home/philipp/facerec/data/at/s23/1.pgm;22
/home/philipp/facerec/data/at/s4/2.pgm;3
/home/philipp/facerec/data/at/s4/7.pgm;3
/home/philipp/facerec/data/at/s4/6.pgm;3
/home/philipp/facerec/data/at/s4/9.pgm;3
/home/philipp/facerec/data/at/s4/5.pgm;3
/home/philipp/facerec/data/at/s4/3.pgm;3
/home/philipp/facerec/data/at/s4/4.pgm;3
/home/philipp/facerec/data/at/s4/10.pgm;3
/home/philipp/facerec/data/at/s4/8.pgm;3
/home/philipp/facerec/data/at/s4/1.pgm;3
/home/philipp/facerec/data/at/s9/2.pgm;8
/home/philipp/facerec/data/at/s9/7.pgm;8
/home/philipp/facerec/data/at/s9/6.pgm;8
/home/philipp/facerec/data/at/s9/9.pgm;8
/home/philipp/facerec/data/at/s9/5.pgm;8
/home/philipp/facerec/data/at/s9/3.pgm;8
/home/philipp/facerec/data/at/s9/4.pgm;8
/home/philipp/facerec/data/at/s9/10.pgm;8
/home/philipp/facerec/data/at/s9/8.pgm;8
/home/philipp/facerec/data/at/s9/1.pgm;8
/home/philipp/facerec/data/at/s37/2.pgm;36
/home/philipp/facerec/data/at/s37/7.pgm;36
/home/philipp/facerec/data/at/s37/6.pgm;36
/home/philipp/facerec/data/at/s37/9.pgm;36
/home/philipp/facerec/data/at/s37/5.pgm;36
/home/philipp/facerec/data/at/s37/3.pgm;36
/home/philipp/facerec/data/at/s37/4.pgm;36
/home/philipp/facerec/data/at/s37/10.pgm;36
/home/philipp/facerec/data/at/s37/8.pgm;36
/home/philipp/facerec/data/at/s37/1.pgm;36
/home/philipp/facerec/data/at/s24/2.pgm;23
/home/philipp/facerec/data/at/s24/7.pgm;23
/home/philipp/facerec/data/at/s24/6.pgm;23
/home/philipp/facerec/data/at/s24/9.pgm;23
/home/philipp/facerec/data/at/s24/5.pgm;23
/home/philipp/facerec/data/at/s24/3.pgm;23
/home/philipp/facerec/data/at/s24/4.pgm;23
/home/philipp/facerec/data/at/s24/10.pgm;23
/home/philipp/facerec/data/at/s24/8.pgm;23
/home/philipp/facerec/data/at/s24/1.pgm;23
/home/philipp/facerec/data/at/s19/2.pgm;18
/home/philipp/facerec/data/at/s19/7.pgm;18
/home/philipp/facerec/data/at/s19/6.pgm;18
/home/philipp/facerec/data/at/s19/9.pgm;18
/home/philipp/facerec/data/at/s19/5.pgm;18
/home/philipp/facerec/data/at/s19/3.pgm;18
/home/philipp/facerec/data/at/s19/4.pgm;18
/home/philipp/facerec/data/at/s19/10.pgm;18
/home/philipp/facerec/data/at/s19/8.pgm;18
/home/philipp/facerec/data/at/s19/1.pgm;18
/home/philipp/facerec/data/at/s8/2.pgm;7
/home/philipp/facerec/data/at/s8/7.pgm;7
/home/philipp/facerec/data/at/s8/6.pgm;7
/home/philipp/facerec/data/at/s8/9.pgm;7
/home/philipp/facerec/data/at/s8/5.pgm;7
/home/philipp/facerec/data/at/s8/3.pgm;7
/home/philipp/facerec/data/at/s8/4.pgm;7
/home/philipp/facerec/data/at/s8/10.pgm;7
/home/philipp/facerec/data/at/s8/8.pgm;7
/home/philipp/facerec/data/at/s8/1.pgm;7
/home/philipp/facerec/data/at/s21/2.pgm;20
/home/philipp/facerec/data/at/s21/7.pgm;20
/home/philipp/facerec/data/at/s21/6.pgm;20
/home/philipp/facerec/data/at/s21/9.pgm;20
/home/philipp/facerec/data/at/s21/5.pgm;20
/home/philipp/facerec/data/at/s21/3.pgm;20
/home/philipp/facerec/data/at/s21/4.pgm;20
/home/philipp/facerec/data/at/s21/10.pgm;20
/home/philipp/facerec/data/at/s21/8.pgm;20
/home/philipp/facerec/data/at/s21/1.pgm;20
/home/philipp/facerec/data/at/s1/2.pgm;0
/home/philipp/facerec/data/at/s1/7.pgm;0
/home/philipp/facerec/data/at/s1/6.pgm;0
/home/philipp/facerec/data/at/s1/9.pgm;0
/home/philipp/facerec/data/at/s1/5.pgm;0
/home/philipp/facerec/data/at/s1/3.pgm;0
/home/philipp/facerec/data/at/s1/4.pgm;0
/home/philipp/facerec/data/at/s1/10.pgm;0
/home/philipp/facerec/data/at/s1/8.pgm;0
/home/philipp/facerec/data/at/s1/1.pgm;0
/home/philipp/facerec/data/at/s7/2.pgm;6
/home/philipp/facerec/data/at/s7/7.pgm;6
/home/philipp/facerec/data/at/s7/6.pgm;6
/home/philipp/facerec/data/at/s7/9.pgm;6
/home/philipp/facerec/data/at/s7/5.pgm;6
/home/philipp/facerec/data/at/s7/3.pgm;6
/home/philipp/facerec/data/at/s7/4.pgm;6
/home/philipp/facerec/data/at/s7/10.pgm;6
/home/philipp/facerec/data/at/s7/8.pgm;6
/home/philipp/facerec/data/at/s7/1.pgm;6
/home/philipp/facerec/data/at/s16/2.pgm;15
/home/philipp/facerec/data/at/s16/7.pgm;15
/home/philipp/facerec/data/at/s16/6.pgm;15
/home/philipp/facerec/data/at/s16/9.pgm;15
/home/philipp/facerec/data/at/s16/5.pgm;15
/home/philipp/facerec/data/at/s16/3.pgm;15
/home/philipp/facerec/data/at/s16/4.pgm;15
/home/philipp/facerec/data/at/s16/10.pgm;15
/home/philipp/facerec/data/at/s16/8.pgm;15
/home/philipp/facerec/data/at/s16/1.pgm;15
/home/philipp/facerec/data/at/s36/2.pgm;35
/home/philipp/facerec/data/at/s36/7.pgm;35
/home/philipp/facerec/data/at/s36/6.pgm;35
/home/philipp/facerec/data/at/s36/9.pgm;35
/home/philipp/facerec/data/at/s36/5.pgm;35
/home/philipp/facerec/data/at/s36/3.pgm;35
/home/philipp/facerec/data/at/s36/4.pgm;35
/home/philipp/facerec/data/at/s36/10.pgm;35
/home/philipp/facerec/data/at/s36/8.pgm;35
/home/philipp/facerec/data/at/s36/1.pgm;35
/home/philipp/facerec/data/at/s25/2.pgm;24
/home/philipp/facerec/data/at/s25/7.pgm;24
/home/philipp/facerec/data/at/s25/6.pgm;24
/home/philipp/facerec/data/at/s25/9.pgm;24
/home/philipp/facerec/data/at/s25/5.pgm;24
/home/philipp/facerec/data/at/s25/3.pgm;24
/home/philipp/facerec/data/at/s25/4.pgm;24
/home/philipp/facerec/data/at/s25/10.pgm;24
/home/philipp/facerec/data/at/s25/8.pgm;24
/home/philipp/facerec/data/at/s25/1.pgm;24
/home/philipp/facerec/data/at/s14/2.pgm;13
/home/philipp/facerec/data/at/s14/7.pgm;13
/home/philipp/facerec/data/at/s14/6.pgm;13
/home/philipp/facerec/data/at/s14/9.pgm;13
/home/philipp/facerec/data/at/s14/5.pgm;13
/home/philipp/facerec/data/at/s14/3.pgm;13
/home/philipp/facerec/data/at/s14/4.pgm;13
/home/philipp/facerec/data/at/s14/10.pgm;13
/home/philipp/facerec/data/at/s14/8.pgm;13
/home/philipp/facerec/data/at/s14/1.pgm;13
/home/philipp/facerec/data/at/s34/2.pgm;33
/home/philipp/facerec/data/at/s34/7.pgm;33
/home/philipp/facerec/data/at/s34/6.pgm;33
/home/philipp/facerec/data/at/s34/9.pgm;33
/home/philipp/facerec/data/at/s34/5.pgm;33
/home/philipp/facerec/data/at/s34/3.pgm;33
/home/philipp/facerec/data/at/s34/4.pgm;33
/home/philipp/facerec/data/at/s34/10.pgm;33
/home/philipp/facerec/data/at/s34/8.pgm;33
/home/philipp/facerec/data/at/s34/1.pgm;33
/home/philipp/facerec/data/at/s11/2.pgm;10
/home/philipp/facerec/data/at/s11/7.pgm;10
/home/philipp/facerec/data/at/s11/6.pgm;10
/home/philipp/facerec/data/at/s11/9.pgm;10
/home/philipp/facerec/data/at/s11/5.pgm;10
/home/philipp/facerec/data/at/s11/3.pgm;10
/home/philipp/facerec/data/at/s11/4.pgm;10
/home/philipp/facerec/data/at/s11/10.pgm;10
/home/philipp/facerec/data/at/s11/8.pgm;10
/home/philipp/facerec/data/at/s11/1.pgm;10
/home/philipp/facerec/data/at/s26/2.pgm;25
/home/philipp/facerec/data/at/s26/7.pgm;25
/home/philipp/facerec/data/at/s26/6.pgm;25
/home/philipp/facerec/data/at/s26/9.pgm;25
/home/philipp/facerec/data/at/s26/5.pgm;25
/home/philipp/facerec/data/at/s26/3.pgm;25
/home/philipp/facerec/data/at/s26/4.pgm;25
/home/philipp/facerec/data/at/s26/10.pgm;25
/home/philipp/facerec/data/at/s26/8.pgm;25
/home/philipp/facerec/data/at/s26/1.pgm;25
/home/philipp/facerec/data/at/s18/2.pgm;17
/home/philipp/facerec/data/at/s18/7.pgm;17
/home/philipp/facerec/data/at/s18/6.pgm;17
/home/philipp/facerec/data/at/s18/9.pgm;17
/home/philipp/facerec/data/at/s18/5.pgm;17
/home/philipp/facerec/data/at/s18/3.pgm;17
/home/philipp/facerec/data/at/s18/4.pgm;17
/home/philipp/facerec/data/at/s18/10.pgm;17
/home/philipp/facerec/data/at/s18/8.pgm;17
/home/philipp/facerec/data/at/s18/1.pgm;17
/home/philipp/facerec/data/at/s29/2.pgm;28
/home/philipp/facerec/data/at/s29/7.pgm;28
/home/philipp/facerec/data/at/s29/6.pgm;28
/home/philipp/facerec/data/at/s29/9.pgm;28
/home/philipp/facerec/data/at/s29/5.pgm;28
/home/philipp/facerec/data/at/s29/3.pgm;28
/home/philipp/facerec/data/at/s29/4.pgm;28
/home/philipp/facerec/data/at/s29/10.pgm;28
/home/philipp/facerec/data/at/s29/8.pgm;28
/home/philipp/facerec/data/at/s29/1.pgm;28
/home/philipp/facerec/data/at/s33/2.pgm;32
/home/philipp/facerec/data/at/s33/7.pgm;32
/home/philipp/facerec/data/at/s33/6.pgm;32
/home/philipp/facerec/data/at/s33/9.pgm;32
/home/philipp/facerec/data/at/s33/5.pgm;32
/home/philipp/facerec/data/at/s33/3.pgm;32
/home/philipp/facerec/data/at/s33/4.pgm;32
/home/philipp/facerec/data/at/s33/10.pgm;32
/home/philipp/facerec/data/at/s33/8.pgm;32
/home/philipp/facerec/data/at/s33/1.pgm;32
/home/philipp/facerec/data/at/s12/2.pgm;11
/home/philipp/facerec/data/at/s12/7.pgm;11
/home/philipp/facerec/data/at/s12/6.pgm;11
/home/philipp/facerec/data/at/s12/9.pgm;11
/home/philipp/facerec/data/at/s12/5.pgm;11
/home/philipp/facerec/data/at/s12/3.pgm;11
/home/philipp/facerec/data/at/s12/4.pgm;11
/home/philipp/facerec/data/at/s12/10.pgm;11
/home/philipp/facerec/data/at/s12/8.pgm;11
/home/philipp/facerec/data/at/s12/1.pgm;11
/home/philipp/facerec/data/at/s6/2.pgm;5
/home/philipp/facerec/data/at/s6/7.pgm;5
/home/philipp/facerec/data/at/s6/6.pgm;5
/home/philipp/facerec/data/at/s6/9.pgm;5
/home/philipp/facerec/data/at/s6/5.pgm;5
/home/philipp/facerec/data/at/s6/3.pgm;5
/home/philipp/facerec/data/at/s6/4.pgm;5
/home/philipp/facerec/data/at/s6/10.pgm;5
/home/philipp/facerec/data/at/s6/8.pgm;5
/home/philipp/facerec/data/at/s6/1.pgm;5
/home/philipp/facerec/data/at/s22/2.pgm;21
/home/philipp/facerec/data/at/s22/7.pgm;21
/home/philipp/facerec/data/at/s22/6.pgm;21
/home/philipp/facerec/data/at/s22/9.pgm;21
/home/philipp/facerec/data/at/s22/5.pgm;21
/home/philipp/facerec/data/at/s22/3.pgm;21
/home/philipp/facerec/data/at/s22/4.pgm;21
/home/philipp/facerec/data/at/s22/10.pgm;21
/home/philipp/facerec/data/at/s22/8.pgm;21
/home/philipp/facerec/data/at/s22/1.pgm;21
/home/philipp/facerec/data/at/s15/2.pgm;14
/home/philipp/facerec/data/at/s15/7.pgm;14
/home/philipp/facerec/data/at/s15/6.pgm;14
/home/philipp/facerec/data/at/s15/9.pgm;14
/home/philipp/facerec/data/at/s15/5.pgm;14
/home/philipp/facerec/data/at/s15/3.pgm;14
/home/philipp/facerec/data/at/s15/4.pgm;14
/home/philipp/facerec/data/at/s15/10.pgm;14
/home/philipp/facerec/data/at/s15/8.pgm;14
/home/philipp/facerec/data/at/s15/1.pgm;14
/home/philipp/facerec/data/at/s2/2.pgm;1
/home/philipp/facerec/data/at/s2/7.pgm;1
/home/philipp/facerec/data/at/s2/6.pgm;1
/home/philipp/facerec/data/at/s2/9.pgm;1
/home/philipp/facerec/data/at/s2/5.pgm;1
/home/philipp/facerec/data/at/s2/3.pgm;1
/home/philipp/facerec/data/at/s2/4.pgm;1
/home/philipp/facerec/data/at/s2/10.pgm;1
/home/philipp/facerec/data/at/s2/8.pgm;1
/home/philipp/facerec/data/at/s2/1.pgm;1
/home/philipp/facerec/data/at/s31/2.pgm;30
/home/philipp/facerec/data/at/s31/7.pgm;30
/home/philipp/facerec/data/at/s31/6.pgm;30
/home/philipp/facerec/data/at/s31/9.pgm;30
/home/philipp/facerec/data/at/s31/5.pgm;30
/home/philipp/facerec/data/at/s31/3.pgm;30
/home/philipp/facerec/data/at/s31/4.pgm;30
/home/philipp/facerec/data/at/s31/10.pgm;30
/home/philipp/facerec/data/at/s31/8.pgm;30
/home/philipp/facerec/data/at/s31/1.pgm;30
/home/philipp/facerec/data/at/s28/2.pgm;27
/home/philipp/facerec/data/at/s28/7.pgm;27
/home/philipp/facerec/data/at/s28/6.pgm;27
/home/philipp/facerec/data/at/s28/9.pgm;27
/home/philipp/facerec/data/at/s28/5.pgm;27
/home/philipp/facerec/data/at/s28/3.pgm;27
/home/philipp/facerec/data/at/s28/4.pgm;27
/home/philipp/facerec/data/at/s28/10.pgm;27
/home/philipp/facerec/data/at/s28/8.pgm;27
/home/philipp/facerec/data/at/s28/1.pgm;27
/home/philipp/facerec/data/at/s40/2.pgm;39
/home/philipp/facerec/data/at/s40/7.pgm;39
/home/philipp/facerec/data/at/s40/6.pgm;39
/home/philipp/facerec/data/at/s40/9.pgm;39
/home/philipp/facerec/data/at/s40/5.pgm;39
/home/philipp/facerec/data/at/s40/3.pgm;39
/home/philipp/facerec/data/at/s40/4.pgm;39
/home/philipp/facerec/data/at/s40/10.pgm;39
/home/philipp/facerec/data/at/s40/8.pgm;39
/home/philipp/facerec/data/at/s40/1.pgm;39
/home/philipp/facerec/data/at/s3/2.pgm;2
/home/philipp/facerec/data/at/s3/7.pgm;2
/home/philipp/facerec/data/at/s3/6.pgm;2
/home/philipp/facerec/data/at/s3/9.pgm;2
/home/philipp/facerec/data/at/s3/5.pgm;2
/home/philipp/facerec/data/at/s3/3.pgm;2
/home/philipp/facerec/data/at/s3/4.pgm;2
/home/philipp/facerec/data/at/s3/10.pgm;2
/home/philipp/facerec/data/at/s3/8.pgm;2
/home/philipp/facerec/data/at/s3/1.pgm;2
/home/philipp/facerec/data/at/s38/2.pgm;37
/home/philipp/facerec/data/at/s38/7.pgm;37
/home/philipp/facerec/data/at/s38/6.pgm;37
/home/philipp/facerec/data/at/s38/9.pgm;37
/home/philipp/facerec/data/at/s38/5.pgm;37
/home/philipp/facerec/data/at/s38/3.pgm;37
/home/philipp/facerec/data/at/s38/4.pgm;37
/home/philipp/facerec/data/at/s38/10.pgm;37
/home/philipp/facerec/data/at/s38/8.pgm;37
/home/philipp/facerec/data/at/s38/1.pgm;37

377
modules/contrib/doc/facerec/facerec_api.rst
View File

@@ -1,377 +0,0 @@
FaceRecognizer
==============
.. highlight:: cpp
.. Sample code::
* An example using the FaceRecognizer class can be found at opencv_source_code/samples/cpp/facerec_demo.cpp
* (Python) An example using the FaceRecognizer class can be found at opencv_source_code/samples/python2/facerec_demo.py
FaceRecognizer
--------------
.. ocv:class:: FaceRecognizer : public Algorithm
All face recognition models in OpenCV are derived from the abstract base class :ocv:class:`FaceRecognizer`, which provides
a unified access to all face recongition algorithms in OpenCV. ::
class FaceRecognizer : public Algorithm
{
public:
//! virtual destructor
virtual ~FaceRecognizer() {}
// Trains a FaceRecognizer.
virtual void train(InputArray src, InputArray labels) = 0;
// Updates a FaceRecognizer.
virtual void update(InputArrayOfArrays src, InputArray labels);
// Gets a prediction from a FaceRecognizer.
virtual int predict(InputArray src) const = 0;
// Predicts the label and confidence for a given sample.
virtual void predict(InputArray src, int &label, double &confidence) const = 0;
// Serializes this object to a given filename.
virtual void save(const String& filename) const;
// Deserializes this object from a given filename.
virtual void load(const String& filename);
// Serializes this object to a given cv::FileStorage.
virtual void save(FileStorage& fs) const = 0;
// Deserializes this object from a given cv::FileStorage.
virtual void load(const FileStorage& fs) = 0;
};
Description
+++++++++++
I'll go a bit more into detail explaining :ocv:class:`FaceRecognizer`, because it doesn't look like a powerful interface at first sight. But: Every :ocv:class:`FaceRecognizer` is an :ocv:class:`Algorithm`, so you can easily get/set all model internals (if allowed by the implementation). :ocv:class:`Algorithm` is a relatively new OpenCV concept, which is available since the 2.4 release. I suggest you take a look at its description.
:ocv:class:`Algorithm` provides the following features for all derived classes:
* So called “virtual constructor”. That is, each Algorithm derivative is registered at program start and you can get the list of registered algorithms and create instance of a particular algorithm by its name (see :ocv:func:`Algorithm::create`). If you plan to add your own algorithms, it is good practice to add a unique prefix to your algorithms to distinguish them from other algorithms.
* Setting/Retrieving algorithm parameters by name. If you used video capturing functionality from OpenCV highgui module, you are probably familar with :ocv:cfunc:`cvSetCaptureProperty`, :ocv:cfunc:`cvGetCaptureProperty`, :ocv:func:`VideoCapture::set` and :ocv:func:`VideoCapture::get`. :ocv:class:`Algorithm` provides similar method where instead of integer id's you specify the parameter names as text Strings. See :ocv:func:`Algorithm::set` and :ocv:func:`Algorithm::get` for details.
* Reading and writing parameters from/to XML or YAML files. Every Algorithm derivative can store all its parameters and then read them back. There is no need to re-implement it each time.
Moreover every :ocv:class:`FaceRecognizer` supports the:
* **Training** of a :ocv:class:`FaceRecognizer` with :ocv:func:`FaceRecognizer::train` on a given set of images (your face database!).
* **Prediction** of a given sample image, that means a face. The image is given as a :ocv:class:`Mat`.
* **Loading/Saving** the model state from/to a given XML or YAML.
.. note:: When using the FaceRecognizer interface in combination with Python, please stick to Python 2. Some underlying scripts like create_csv will not work in other versions, like Python 3.
Setting the Thresholds
+++++++++++++++++++++++
Sometimes you run into the situation, when you want to apply a threshold on the prediction. A common scenario in face recognition is to tell, whether a face belongs to the training dataset or if it is unknown. You might wonder, why there's no public API in :ocv:class:`FaceRecognizer` to set the threshold for the prediction, but rest assured: It's supported. It just means there's no generic way in an abstract class to provide an interface for setting/getting the thresholds of *every possible* :ocv:class:`FaceRecognizer` algorithm. The appropriate place to set the thresholds is in the constructor of the specific :ocv:class:`FaceRecognizer` and since every :ocv:class:`FaceRecognizer` is a :ocv:class:`Algorithm` (see above), you can get/set the thresholds at runtime!
Here is an example of setting a threshold for the Eigenfaces method, when creating the model:
.. code-block:: cpp
// Let's say we want to keep 10 Eigenfaces and have a threshold value of 10.0
int num_components = 10;
double threshold = 10.0;
// Then if you want to have a cv::FaceRecognizer with a confidence threshold,
// create the concrete implementation with the appropiate parameters:
Ptr<FaceRecognizer> model = createEigenFaceRecognizer(num_components, threshold);
Sometimes it's impossible to train the model, just to experiment with threshold values. Thanks to :ocv:class:`Algorithm` it's possible to set internal model thresholds during runtime. Let's see how we would set/get the prediction for the Eigenface model, we've created above:
.. code-block:: cpp
// The following line reads the threshold from the Eigenfaces model:
double current_threshold = model->getDouble("threshold");
// And this line sets the threshold to 0.0:
model->set("threshold", 0.0);
If you've set the threshold to ``0.0`` as we did above, then:
.. code-block:: cpp
//
Mat img = imread("person1/3.jpg", CV_LOAD_IMAGE_GRAYSCALE);
// Get a prediction from the model. Note: We've set a threshold of 0.0 above,
// since the distance is almost always larger than 0.0, you'll get -1 as
// label, which indicates, this face is unknown
int predicted_label = model->predict(img);
// ...
is going to yield ``-1`` as predicted label, which states this face is unknown.
Getting the name of a FaceRecognizer
+++++++++++++++++++++++++++++++++++++
Since every :ocv:class:`FaceRecognizer` is a :ocv:class:`Algorithm`, you can use :ocv:func:`Algorithm::name` to get the name of a :ocv:class:`FaceRecognizer`:
.. code-block:: cpp
// Create a FaceRecognizer:
Ptr<FaceRecognizer> model = createEigenFaceRecognizer();
// And here's how to get its name:
String name = model->name();
FaceRecognizer::train
---------------------
Trains a FaceRecognizer with given data and associated labels.
.. ocv:function:: void FaceRecognizer::train( InputArrayOfArrays src, InputArray labels ) = 0
:param src: The training images, that means the faces you want to learn. The data has to be given as a ``vector<Mat>``.
:param labels: The labels corresponding to the images have to be given either as a ``vector<int>`` or a
The following source code snippet shows you how to learn a Fisherfaces model on a given set of images. The images are read with :ocv:func:`imread` and pushed into a ``std::vector<Mat>``. The labels of each image are stored within a ``std::vector<int>`` (you could also use a :ocv:class:`Mat` of type `CV_32SC1`). Think of the label as the subject (the person) this image belongs to, so same subjects (persons) should have the same label. For the available :ocv:class:`FaceRecognizer` you don't have to pay any attention to the order of the labels, just make sure same persons have the same label:
.. code-block:: cpp
// holds images and labels
vector<Mat> images;
vector<int> labels;
// images for first person
images.push_back(imread("person0/0.jpg", CV_LOAD_IMAGE_GRAYSCALE)); labels.push_back(0);
images.push_back(imread("person0/1.jpg", CV_LOAD_IMAGE_GRAYSCALE)); labels.push_back(0);
images.push_back(imread("person0/2.jpg", CV_LOAD_IMAGE_GRAYSCALE)); labels.push_back(0);
// images for second person
images.push_back(imread("person1/0.jpg", CV_LOAD_IMAGE_GRAYSCALE)); labels.push_back(1);
images.push_back(imread("person1/1.jpg", CV_LOAD_IMAGE_GRAYSCALE)); labels.push_back(1);
images.push_back(imread("person1/2.jpg", CV_LOAD_IMAGE_GRAYSCALE)); labels.push_back(1);
Now that you have read some images, we can create a new :ocv:class:`FaceRecognizer`. In this example I'll create a Fisherfaces model and decide to keep all of the possible Fisherfaces:
.. code-block:: cpp
// Create a new Fisherfaces model and retain all available Fisherfaces,
// this is the most common usage of this specific FaceRecognizer:
//
Ptr<FaceRecognizer> model = createFisherFaceRecognizer();
And finally train it on the given dataset (the face images and labels):
.. code-block:: cpp
// This is the common interface to train all of the available cv::FaceRecognizer
// implementations:
//
model->train(images, labels);
FaceRecognizer::update
----------------------
Updates a FaceRecognizer with given data and associated labels.
.. ocv:function:: void FaceRecognizer::update( InputArrayOfArrays src, InputArray labels )
:param src: The training images, that means the faces you want to learn. The data has to be given as a ``vector<Mat>``.
:param labels: The labels corresponding to the images have to be given either as a ``vector<int>`` or a
This method updates a (probably trained) :ocv:class:`FaceRecognizer`, but only if the algorithm supports it. The Local Binary Patterns Histograms (LBPH) recognizer (see :ocv:func:`createLBPHFaceRecognizer`) can be updated. For the Eigenfaces and Fisherfaces method, this is algorithmically not possible and you have to re-estimate the model with :ocv:func:`FaceRecognizer::train`. In any case, a call to train empties the existing model and learns a new model, while update does not delete any model data.
.. code-block:: cpp
// Create a new LBPH model (it can be updated) and use the default parameters,
// this is the most common usage of this specific FaceRecognizer:
//
Ptr<FaceRecognizer> model = createLBPHFaceRecognizer();
// This is the common interface to train all of the available cv::FaceRecognizer
// implementations:
//
model->train(images, labels);
// Some containers to hold new image:
vector<Mat> newImages;
vector<int> newLabels;
// You should add some images to the containers:
//
// ...
//
// Now updating the model is as easy as calling:
model->update(newImages,newLabels);
// This will preserve the old model data and extend the existing model
// with the new features extracted from newImages!
Calling update on an Eigenfaces model (see :ocv:func:`createEigenFaceRecognizer`), which doesn't support updating, will throw an error similar to:
.. code-block:: none
OpenCV Error: The function/feature is not implemented (This FaceRecognizer (FaceRecognizer.Eigenfaces) does not support updating, you have to use FaceRecognizer::train to update it.) in update, file /home/philipp/git/opencv/modules/contrib/src/facerec.cpp, line 305
terminate called after throwing an instance of 'cv::Exception'
Please note: The :ocv:class:`FaceRecognizer` does not store your training images, because this would be very memory intense and it's not the responsibility of te :ocv:class:`FaceRecognizer` to do so. The caller is responsible for maintaining the dataset, he want to work with.
FaceRecognizer::predict
-----------------------
.. ocv:function:: int FaceRecognizer::predict( InputArray src ) const = 0
.. ocv:function:: void FaceRecognizer::predict( InputArray src, int & label, double & confidence ) const = 0
Predicts a label and associated confidence (e.g. distance) for a given input image.
:param src: Sample image to get a prediction from.
:param label: The predicted label for the given image.
:param confidence: Associated confidence (e.g. distance) for the predicted label.
The suffix ``const`` means that prediction does not affect the internal model
state, so the method can be safely called from within different threads.
The following example shows how to get a prediction from a trained model:
.. code-block:: cpp
using namespace cv;
// Do your initialization here (create the cv::FaceRecognizer model) ...
// ...
// Read in a sample image:
Mat img = imread("person1/3.jpg", CV_LOAD_IMAGE_GRAYSCALE);
// And get a prediction from the cv::FaceRecognizer:
int predicted = model->predict(img);
Or to get a prediction and the associated confidence (e.g. distance):
.. code-block:: cpp
using namespace cv;
// Do your initialization here (create the cv::FaceRecognizer model) ...
// ...
Mat img = imread("person1/3.jpg", CV_LOAD_IMAGE_GRAYSCALE);
// Some variables for the predicted label and associated confidence (e.g. distance):
int predicted_label = -1;
double predicted_confidence = 0.0;
// Get the prediction and associated confidence from the model
model->predict(img, predicted_label, predicted_confidence);
FaceRecognizer::save
--------------------
Saves a :ocv:class:`FaceRecognizer` and its model state.
.. ocv:function:: void FaceRecognizer::save(const String& filename) const
Saves this model to a given filename, either as XML or YAML.
:param filename: The filename to store this :ocv:class:`FaceRecognizer` to (either XML/YAML).
.. ocv:function:: void FaceRecognizer::save(FileStorage& fs) const
Saves this model to a given :ocv:class:`FileStorage`.
:param fs: The :ocv:class:`FileStorage` to store this :ocv:class:`FaceRecognizer` to.
Every :ocv:class:`FaceRecognizer` overwrites ``FaceRecognizer::save(FileStorage& fs)``
to save the internal model state. ``FaceRecognizer::save(const String& filename)`` saves
the state of a model to the given filename.
The suffix ``const`` means that prediction does not affect the internal model
state, so the method can be safely called from within different threads.
FaceRecognizer::load
--------------------
Loads a :ocv:class:`FaceRecognizer` and its model state.
.. ocv:function:: void FaceRecognizer::load( const String& filename )
.. ocv:function:: void FaceRecognizer::load( const FileStorage& fs ) = 0
Loads a persisted model and state from a given XML or YAML file . Every
:ocv:class:`FaceRecognizer` has to overwrite ``FaceRecognizer::load(FileStorage& fs)``
to enable loading the model state. ``FaceRecognizer::load(FileStorage& fs)`` in
turn gets called by ``FaceRecognizer::load(const String& filename)``, to ease
saving a model.
createEigenFaceRecognizer
-------------------------
.. ocv:function:: Ptr<FaceRecognizer> createEigenFaceRecognizer(int num_components = 0, double threshold = DBL_MAX)
:param num_components: The number of components (read: Eigenfaces) kept for this Prinicpal Component Analysis. As a hint: There's no rule how many components (read: Eigenfaces) should be kept for good reconstruction capabilities. It is based on your input data, so experiment with the number. Keeping 80 components should almost always be sufficient.
:param threshold: The threshold applied in the prediciton.
Notes:
++++++
* Training and prediction must be done on grayscale images, use :ocv:func:`cvtColor` to convert between the color spaces.
* **THE EIGENFACES METHOD MAKES THE ASSUMPTION, THAT THE TRAINING AND TEST IMAGES ARE OF EQUAL SIZE.** (caps-lock, because I got so many mails asking for this). You have to make sure your input data has the correct shape, else a meaningful exception is thrown. Use :ocv:func:`resize` to resize the images.
* This model does not support updating.
Model internal data:
++++++++++++++++++++
* ``num_components`` see :ocv:func:`createEigenFaceRecognizer`.
* ``threshold`` see :ocv:func:`createEigenFaceRecognizer`.
* ``eigenvalues`` The eigenvalues for this Principal Component Analysis (ordered descending).
* ``eigenvectors`` The eigenvectors for this Principal Component Analysis (ordered by their eigenvalue).
* ``mean`` The sample mean calculated from the training data.
* ``projections`` The projections of the training data.
* ``labels`` The threshold applied in the prediction. If the distance to the nearest neighbor is larger than the threshold, this method returns -1.
createFisherFaceRecognizer
--------------------------
.. ocv:function:: Ptr<FaceRecognizer> createFisherFaceRecognizer(int num_components = 0, double threshold = DBL_MAX)
:param num_components: The number of components (read: Fisherfaces) kept for this Linear Discriminant Analysis with the Fisherfaces criterion. It's useful to keep all components, that means the number of your classes ``c`` (read: subjects, persons you want to recognize). If you leave this at the default (``0``) or set it to a value less-equal ``0`` or greater ``(c-1)``, it will be set to the correct number ``(c-1)`` automatically.
:param threshold: The threshold applied in the prediction. If the distance to the nearest neighbor is larger than the threshold, this method returns -1.
Notes:
++++++
* Training and prediction must be done on grayscale images, use :ocv:func:`cvtColor` to convert between the color spaces.
* **THE FISHERFACES METHOD MAKES THE ASSUMPTION, THAT THE TRAINING AND TEST IMAGES ARE OF EQUAL SIZE.** (caps-lock, because I got so many mails asking for this). You have to make sure your input data has the correct shape, else a meaningful exception is thrown. Use :ocv:func:`resize` to resize the images.
* This model does not support updating.
Model internal data:
++++++++++++++++++++
* ``num_components`` see :ocv:func:`createFisherFaceRecognizer`.
* ``threshold`` see :ocv:func:`createFisherFaceRecognizer`.
* ``eigenvalues`` The eigenvalues for this Linear Discriminant Analysis (ordered descending).
* ``eigenvectors`` The eigenvectors for this Linear Discriminant Analysis (ordered by their eigenvalue).
* ``mean`` The sample mean calculated from the training data.
* ``projections`` The projections of the training data.
* ``labels`` The labels corresponding to the projections.
createLBPHFaceRecognizer
-------------------------
.. ocv:function:: Ptr<FaceRecognizer> createLBPHFaceRecognizer(int radius=1, int neighbors=8, int grid_x=8, int grid_y=8, double threshold = DBL_MAX)
:param radius: The radius used for building the Circular Local Binary Pattern. The greater the radius, the
:param neighbors: The number of sample points to build a Circular Local Binary Pattern from. An appropriate value is to use `` 8`` sample points. Keep in mind: the more sample points you include, the higher the computational cost.
:param grid_x: The number of cells in the horizontal direction, ``8`` is a common value used in publications. The more cells, the finer the grid, the higher the dimensionality of the resulting feature vector.
:param grid_y: The number of cells in the vertical direction, ``8`` is a common value used in publications. The more cells, the finer the grid, the higher the dimensionality of the resulting feature vector.
:param threshold: The threshold applied in the prediction. If the distance to the nearest neighbor is larger than the threshold, this method returns -1.
Notes:
++++++
* The Circular Local Binary Patterns (used in training and prediction) expect the data given as grayscale images, use :ocv:func:`cvtColor` to convert between the color spaces.
* This model supports updating.
Model internal data:
++++++++++++++++++++
* ``radius`` see :ocv:func:`createLBPHFaceRecognizer`.
* ``neighbors`` see :ocv:func:`createLBPHFaceRecognizer`.
* ``grid_x`` see :ocv:func:`createLBPHFaceRecognizer`.
* ``grid_y`` see :ocv:func:`createLBPHFaceRecognizer`.
* ``threshold`` see :ocv:func:`createLBPHFaceRecognizer`.
* ``histograms`` Local Binary Patterns Histograms calculated from the given training data (empty if none was given).
* ``labels`` Labels corresponding to the calculated Local Binary Patterns Histograms.

86
modules/contrib/doc/facerec/facerec_changelog.rst
View File

@@ -1,86 +0,0 @@
Changelog
=========
Release 0.05
------------
This library is now included in the official OpenCV distribution (from 2.4 on).
The :ocv:class`FaceRecognizer` is now an :ocv:class:`Algorithm`, which better fits into the overall
OpenCV API.
To reduce the confusion on user side and minimize my work, libfacerec and OpenCV
have been synchronized and are now based on the same interfaces and implementation.
The library now has an extensive documentation:
* The API is explained in detail and with a lot of code examples.
* The face recognition guide I had written for Python and GNU Octave/MATLAB has been adapted to the new OpenCV C++ ``cv::FaceRecognizer``.
* A tutorial for gender classification with Fisherfaces.
* A tutorial for face recognition in videos (e.g. webcam).
Release highlights
++++++++++++++++++
* There are no single highlights to pick from, this release is a highlight itself.
Release 0.04
------------
This version is fully Windows-compatible and works with OpenCV 2.3.1. Several
bugfixes, but none influenced the recognition rate.
Release highlights
++++++++++++++++++
* A whole lot of exceptions with meaningful error messages.
* A tutorial for Windows users: `http://bytefish.de/blog/opencv_visual_studio_and_libfacerec <http://bytefish.de/blog/opencv_visual_studio_and_libfacerec>`_
Release 0.03
------------
Reworked the library to provide separate implementations in cpp files, because
it's the preferred way of contributing OpenCV libraries. This means the library
is not header-only anymore. Slight API changes were done, please see the
documentation for details.
Release highlights
++++++++++++++++++
* New Unit Tests (for LBP Histograms) make the library more robust.
* Added more documentation.
Release 0.02
------------
Reworked the library to provide separate implementations in cpp files, because
it's the preferred way of contributing OpenCV libraries. This means the library
is not header-only anymore. Slight API changes were done, please see the
documentation for details.
Release highlights
++++++++++++++++++
* New Unit Tests (for LBP Histograms) make the library more robust.
* Added a documentation and changelog in reStructuredText.
Release 0.01
------------
Initial release as header-only library.
Release highlights
++++++++++++++++++
* Colormaps for OpenCV to enhance the visualization.
* Face Recognition algorithms implemented:
* Eigenfaces [TP91]_
* Fisherfaces [BHK97]_
* Local Binary Patterns Histograms [AHP04]_
* Added persistence facilities to store the models with a common API.
* Unit Tests (using `gtest <http://code.google.com/p/googletest/>`_).
* Providing a CMakeLists.txt to enable easy cross-platform building.

628
modules/contrib/doc/facerec/facerec_tutorial.rst
View File

@@ -1,628 +0,0 @@
Face Recognition with OpenCV
############################
.. contents:: Table of Contents
:depth: 3
Introduction
============
`OpenCV (Open Source Computer Vision) <http://opencv.org>`_ is a popular computer vision library started by `Intel <http://www.intel.com>`_ in 1999. The cross-platform library sets its focus on real-time image processing and includes patent-free implementations of the latest computer vision algorithms. In 2008 `Willow Garage <http://www.willowgarage.com>`_ took over support and OpenCV 2.3.1 now comes with a programming interface to C, C++, `Python <http://www.python.org>`_ and `Android <http://www.android.com>`_. OpenCV is released under a BSD license so it is used in academic projects and commercial products alike.
OpenCV 2.4 now comes with the very new :ocv:class:`FaceRecognizer` class for face recognition, so you can start experimenting with face recognition right away. This document is the guide I've wished for, when I was working myself into face recognition. It shows you how to perform face recognition with :ocv:class:`FaceRecognizer` in OpenCV (with full source code listings) and gives you an introduction into the algorithms behind. I'll also show how to create the visualizations you can find in many publications, because a lot of people asked for.
The currently available algorithms are:
* Eigenfaces (see :ocv:func:`createEigenFaceRecognizer`)
* Fisherfaces (see :ocv:func:`createFisherFaceRecognizer`)
* Local Binary Patterns Histograms (see :ocv:func:`createLBPHFaceRecognizer`)
You don't need to copy and paste the source code examples from this page, because they are available in the ``src`` folder coming with this documentation. If you have built OpenCV with the samples turned on, chances are good you have them compiled already! Although it might be interesting for very advanced users, I've decided to leave the implementation details out as I am afraid they confuse new users.
All code in this document is released under the `BSD license <http://www.opensource.org/licenses/bsd-license>`_, so feel free to use it for your projects.
Face Recognition
================
Face recognition is an easy task for humans. Experiments in [Tu06]_ have shown, that even one to three day old babies are able to distinguish between known faces. So how hard could it be for a computer? It turns out we know little about human recognition to date. Are inner features (eyes, nose, mouth) or outer features (head shape, hairline) used for a successful face recognition? How do we analyze an image and how does the brain encode it? It was shown by `David Hubel <http://en.wikipedia.org/wiki/David_H._Hubel>`_ and `Torsten Wiesel <http://en.wikipedia.org/wiki/Torsten_Wiesel>`_, that our brain has specialized nerve cells responding to specific local features of a scene, such as lines, edges, angles or movement. Since we don't see the world as scattered pieces, our visual cortex must somehow combine the different sources of information into useful patterns. Automatic face recognition is all about extracting those meaningful features from an image, putting them into a useful representation and performing some kind of classification on them.
Face recognition based on the geometric features of a face is probably the most intuitive approach to face recognition. One of the first automated face recognition systems was described in [Kanade73]_: marker points (position of eyes, ears, nose, ...) were used to build a feature vector (distance between the points, angle between them, ...). The recognition was performed by calculating the euclidean distance between feature vectors of a probe and reference image. Such a method is robust against changes in illumination by its nature, but has a huge drawback: the accurate registration of the marker points is complicated, even with state of the art algorithms. Some of the latest work on geometric face recognition was carried out in [Bru92]_. A 22-dimensional feature vector was used and experiments on large datasets have shown, that geometrical features alone my not carry enough information for face recognition.
The Eigenfaces method described in [TP91]_ took a holistic approach to face recognition: A facial image is a point from a high-dimensional image space and a lower-dimensional representation is found, where classification becomes easy. The lower-dimensional subspace is found with Principal Component Analysis, which identifies the axes with maximum variance. While this kind of transformation is optimal from a reconstruction standpoint, it doesn't take any class labels into account. Imagine a situation where the variance is generated from external sources, let it be light. The axes with maximum variance do not necessarily contain any discriminative information at all, hence a classification becomes impossible. So a class-specific projection with a Linear Discriminant Analysis was applied to face recognition in [BHK97]_. The basic idea is to minimize the variance within a class, while maximizing the variance between the classes at the same time.
Recently various methods for a local feature extraction emerged. To avoid the high-dimensionality of the input data only local regions of an image are described, the extracted features are (hopefully) more robust against partial occlusion, illumation and small sample size. Algorithms used for a local feature extraction are Gabor Wavelets ([Wiskott97]_), Discrete Cosinus Transform ([Messer06]_) and Local Binary Patterns ([AHP04]_). It's still an open research question what's the best way to preserve spatial information when applying a local feature extraction, because spatial information is potentially useful information.
Face Database
==============
Let's get some data to experiment with first. I don't want to do a toy example here. We are doing face recognition, so you'll need some face images! You can either create your own dataset or start with one of the available face databases, `http://face-rec.org/databases/ <http://face-rec.org/databases>`_ gives you an up-to-date overview. Three interesting databases are (parts of the description are quoted from `http://face-rec.org <http://face-rec.org>`_):
* `AT&T Facedatabase <http://www.cl.cam.ac.uk/research/dtg/attarchive/facedatabase.html>`_ The AT&T Facedatabase, sometimes also referred to as *ORL Database of Faces*, contains ten different images of each of 40 distinct subjects. For some subjects, the images were taken at different times, varying the lighting, facial expressions (open / closed eyes, smiling / not smiling) and facial details (glasses / no glasses). All the images were taken against a dark homogeneous background with the subjects in an upright, frontal position (with tolerance for some side movement).
* `Yale Facedatabase A <http://vision.ucsd.edu/content/yale-face-database>`_, also known as Yalefaces. The AT&T Facedatabase is good for initial tests, but it's a fairly easy database. The Eigenfaces method already has a 97% recognition rate on it, so you won't see any great improvements with other algorithms. The Yale Facedatabase A (also known as Yalefaces) is a more appropriate dataset for initial experiments, because the recognition problem is harder. The database consists of 15 people (14 male, 1 female) each with 11 grayscale images sized :math:`320 \times 243` pixel. There are changes in the light conditions (center light, left light, right light), facial expressions (happy, normal, sad, sleepy, surprised, wink) and glasses (glasses, no-glasses).
The original images are not cropped and aligned. Please look into the :ref:`appendixft` for a Python script, that does the job for you.
* `Extended Yale Facedatabase B <http://vision.ucsd.edu/~leekc/ExtYaleDatabase/ExtYaleB.html>`_ The Extended Yale Facedatabase B contains 2414 images of 38 different people in its cropped version. The focus of this database is set on extracting features that are robust to illumination, the images have almost no variation in emotion/occlusion/... . I personally think, that this dataset is too large for the experiments I perform in this document. You better use the `AT&T Facedatabase <http://www.cl.cam.ac.uk/research/dtg/attarchive/facedatabase.html>`_ for intial testing. A first version of the Yale Facedatabase B was used in [BHK97]_ to see how the Eigenfaces and Fisherfaces method perform under heavy illumination changes. [Lee05]_ used the same setup to take 16128 images of 28 people. The Extended Yale Facedatabase B is the merge of the two databases, which is now known as Extended Yalefacedatabase B.
Preparing the data
-------------------
Once we have acquired some data, we'll need to read it in our program. In the demo applications I have decided to read the images from a very simple CSV file. Why? Because it's the simplest platform-independent approach I can think of. However, if you know a simpler solution please ping me about it. Basically all the CSV file needs to contain are lines composed of a ``filename`` followed by a ``;`` followed by the ``label`` (as *integer number*), making up a line like this:
.. code-block:: none
/path/to/image.ext;0
Let's dissect the line. ``/path/to/image.ext`` is the path to an image, probably something like this if you are in Windows: ``C:/faces/person0/image0.jpg``. Then there is the separator ``;`` and finally we assign the label ``0`` to the image. Think of the label as the subject (the person) this image belongs to, so same subjects (persons) should have the same label.
Download the AT&T Facedatabase from AT&T Facedatabase and the corresponding CSV file from at.txt, which looks like this (file is without ... of course):
.. code-block:: none
./at/s1/1.pgm;0
./at/s1/2.pgm;0
...
./at/s2/1.pgm;1
./at/s2/2.pgm;1
...
./at/s40/1.pgm;39
./at/s40/2.pgm;39
Imagine I have extracted the files to ``D:/data/at`` and have downloaded the CSV file to ``D:/data/at.txt``. Then you would simply need to Search & Replace ``./`` with ``D:/data/``. You can do that in an editor of your choice, every sufficiently advanced editor can do this. Once you have a CSV file with valid filenames and labels, you can run any of the demos by passing the path to the CSV file as parameter:
.. code-block:: none
facerec_demo.exe D:/data/at.txt
Creating the CSV File
+++++++++++++++++++++
You don't really want to create the CSV file by hand. I have prepared you a little Python script ``create_csv.py`` (you find it at ``src/create_csv.py`` coming with this tutorial) that automatically creates you a CSV file. If you have your images in hierarchie like this (``/basepath/<subject>/<image.ext>``):
.. code-block:: none
philipp@mango:~/facerec/data/at$ tree
.
|-- s1
| |-- 1.pgm
| |-- ...
| |-- 10.pgm
|-- s2
| |-- 1.pgm
| |-- ...
| |-- 10.pgm
...
|-- s40
| |-- 1.pgm
| |-- ...
| |-- 10.pgm
Then simply call create_csv.py with the path to the folder, just like this and you could save the output:
.. code-block:: none
philipp@mango:~/facerec/data$ python create_csv.py
at/s13/2.pgm;0
at/s13/7.pgm;0
at/s13/6.pgm;0
at/s13/9.pgm;0
at/s13/5.pgm;0
at/s13/3.pgm;0
at/s13/4.pgm;0
at/s13/10.pgm;0
at/s13/8.pgm;0
at/s13/1.pgm;0
at/s17/2.pgm;1
at/s17/7.pgm;1
at/s17/6.pgm;1
at/s17/9.pgm;1
at/s17/5.pgm;1
at/s17/3.pgm;1
[...]
Please see the :ref:`appendixft` for additional informations.
Eigenfaces
==========
The problem with the image representation we are given is its high dimensionality. Two-dimensional :math:`p \times q` grayscale images span a :math:`m = pq`-dimensional vector space, so an image with :math:`100 \times 100` pixels lies in a :math:`10,000`-dimensional image space already. The question is: Are all dimensions equally useful for us? We can only make a decision if there's any variance in data, so what we are looking for are the components that account for most of the information. The Principal Component Analysis (PCA) was independently proposed by `Karl Pearson <http://en.wikipedia.org/wiki/Karl_Pearson>`_ (1901) and `Harold Hotelling <http://en.wikipedia.org/wiki/Harold_Hotelling>`_ (1933) to turn a set of possibly correlated variables into a smaller set of uncorrelated variables. The idea is, that a high-dimensional dataset is often described by correlated variables and therefore only a few meaningful dimensions account for most of the information. The PCA method finds the directions with the greatest variance in the data, called principal components.
Algorithmic Description
-----------------------
Let :math:`X = \{ x_{1}, x_{2}, \ldots, x_{n} \}` be a random vector with observations :math:`x_i \in R^{d}`.
1. Compute the mean :math:`\mu`
.. math::
\mu = \frac{1}{n} \sum_{i=1}^{n} x_{i}
2. Compute the the Covariance Matrix `S`
.. math::
S = \frac{1}{n} \sum_{i=1}^{n} (x_{i} - \mu) (x_{i} - \mu)^{T}`
3. Compute the eigenvalues :math:`\lambda_{i}` and eigenvectors :math:`v_{i}` of :math:`S`
.. math::
S v_{i} = \lambda_{i} v_{i}, i=1,2,\ldots,n
4. Order the eigenvectors descending by their eigenvalue. The :math:`k` principal components are the eigenvectors corresponding to the :math:`k` largest eigenvalues.
The :math:`k` principal components of the observed vector :math:`x` are then given by:
.. math::
y = W^{T} (x - \mu)
where :math:`W = (v_{1}, v_{2}, \ldots, v_{k})`.
The reconstruction from the PCA basis is given by:
.. math::
x = W y + \mu
where :math:`W = (v_{1}, v_{2}, \ldots, v_{k})`.
The Eigenfaces method then performs face recognition by:
* Projecting all training samples into the PCA subspace.
* Projecting the query image into the PCA subspace.
* Finding the nearest neighbor between the projected training images and the projected query image.
Still there's one problem left to solve. Imagine we are given :math:`400` images sized :math:`100 \times 100` pixel. The Principal Component Analysis solves the covariance matrix :math:`S = X X^{T}`, where :math:`{size}(X) = 10000 \times 400` in our example. You would end up with a :math:`10000 \times 10000` matrix, roughly :math:`0.8 GB`. Solving this problem isn't feasible, so we'll need to apply a trick. From your linear algebra lessons you know that a :math:`M \times N` matrix with :math:`M > N` can only have :math:`N - 1` non-zero eigenvalues. So it's possible to take the eigenvalue decomposition :math:`S = X^{T} X` of size :math:`N \times N` instead:
.. math::
X^{T} X v_{i} = \lambda_{i} v{i}
and get the original eigenvectors of :math:`S = X X^{T}` with a left multiplication of the data matrix:
.. math::
X X^{T} (X v_{i}) = \lambda_{i} (X v_{i})
The resulting eigenvectors are orthogonal, to get orthonormal eigenvectors they need to be normalized to unit length. I don't want to turn this into a publication, so please look into [Duda01]_ for the derivation and proof of the equations.
Eigenfaces in OpenCV
--------------------
For the first source code example, I'll go through it with you. I am first giving you the whole source code listing, and after this we'll look at the most important lines in detail. Please note: every source code listing is commented in detail, so you should have no problems following it.
.. literalinclude:: src/facerec_eigenfaces.cpp
:language: cpp
:linenos:
The source code for this demo application is also available in the ``src`` folder coming with this documentation:
* :download:`src/facerec_eigenfaces.cpp <src/facerec_eigenfaces.cpp>`
I've used the jet colormap, so you can see how the grayscale values are distributed within the specific Eigenfaces. You can see, that the Eigenfaces do not only encode facial features, but also the illumination in the images (see the left light in Eigenface \#4, right light in Eigenfaces \#5):
.. image:: img/eigenfaces_opencv.png
:align: center
We've already seen, that we can reconstruct a face from its lower dimensional approximation. So let's see how many Eigenfaces are needed for a good reconstruction. I'll do a subplot with :math:`10,30,\ldots,310` Eigenfaces:
.. code-block:: cpp
// Display or save the image reconstruction at some predefined steps:
for(int num_components = 10; num_components < 300; num_components+=15) {
// slice the eigenvectors from the model
Mat evs = Mat(W, Range::all(), Range(0, num_components));
Mat projection = subspaceProject(evs, mean, images[0].reshape(1,1));
Mat reconstruction = subspaceReconstruct(evs, mean, projection);
// Normalize the result:
reconstruction = norm_0_255(reconstruction.reshape(1, images[0].rows));
// Display or save:
if(argc == 2) {
imshow(format("eigenface_reconstruction_%d", num_components), reconstruction);
} else {
imwrite(format("%s/eigenface_reconstruction_%d.png", output_folder.c_str(), num_components), reconstruction);
}
}
10 Eigenvectors are obviously not sufficient for a good image reconstruction, 50 Eigenvectors may already be sufficient to encode important facial features. You'll get a good reconstruction with approximately 300 Eigenvectors for the AT&T Facedatabase. There are rule of thumbs how many Eigenfaces you should choose for a successful face recognition, but it heavily depends on the input data. [Zhao03]_ is the perfect point to start researching for this:
.. image:: img/eigenface_reconstruction_opencv.png
:align: center
Fisherfaces
============
The Principal Component Analysis (PCA), which is the core of the Eigenfaces method, finds a linear combination of features that maximizes the total variance in data. While this is clearly a powerful way to represent data, it doesn't consider any classes and so a lot of discriminative information *may* be lost when throwing components away. Imagine a situation where the variance in your data is generated by an external source, let it be the light. The components identified by a PCA do not necessarily contain any discriminative information at all, so the projected samples are smeared together and a classification becomes impossible (see `http://www.bytefish.de/wiki/pca_lda_with_gnu_octave <http://www.bytefish.de/wiki/pca_lda_with_gnu_octave>`_ for an example).
The Linear Discriminant Analysis performs a class-specific dimensionality reduction and was invented by the great statistician `Sir R. A. Fisher <http://en.wikipedia.org/wiki/Ronald_Fisher>`_. He successfully used it for classifying flowers in his 1936 paper *The use of multiple measurements in taxonomic problems* [Fisher36]_. In order to find the combination of features that separates best between classes the Linear Discriminant Analysis maximizes the ratio of between-classes to within-classes scatter, instead of maximizing the overall scatter. The idea is simple: same classes should cluster tightly together, while different classes are as far away as possible from each other in the lower-dimensional representation. This was also recognized by `Belhumeur <http://www.cs.columbia.edu/~belhumeur/>`_, `Hespanha <http://www.ece.ucsb.edu/~hespanha/>`_ and `Kriegman <http://cseweb.ucsd.edu/~kriegman/>`_ and so they applied a Discriminant Analysis to face recognition in [BHK97]_.
Algorithmic Description
-----------------------
Let :math:`X` be a random vector with samples drawn from :math:`c` classes:
.. math::
:nowrap:
\begin{align*}
X & = & \{X_1,X_2,\ldots,X_c\} \\
X_i & = & \{x_1, x_2, \ldots, x_n\}
\end{align*}
The scatter matrices :math:`S_{B}` and `S_{W}` are calculated as:
.. math::
:nowrap:
\begin{align*}
S_{B} & = & \sum_{i=1}^{c} N_{i} (\mu_i - \mu)(\mu_i - \mu)^{T} \\
S_{W} & = & \sum_{i=1}^{c} \sum_{x_{j} \in X_{i}} (x_j - \mu_i)(x_j - \mu_i)^{T}
\end{align*}
, where :math:`\mu` is the total mean:
.. math::
\mu = \frac{1}{N} \sum_{i=1}^{N} x_i
And :math:`\mu_i` is the mean of class :math:`i \in \{1,\ldots,c\}`:
.. math::
\mu_i = \frac{1}{|X_i|} \sum_{x_j \in X_i} x_j
Fisher's classic algorithm now looks for a projection :math:`W`, that maximizes the class separability criterion:
.. math::
W_{opt} = \operatorname{arg\,max}_{W} \frac{|W^T S_B W|}{|W^T S_W W|}
Following [BHK97]_, a solution for this optimization problem is given by solving the General Eigenvalue Problem:
.. math::
:nowrap:
\begin{align*}
S_{B} v_{i} & = & \lambda_{i} S_w v_{i} \nonumber \\
S_{W}^{-1} S_{B} v_{i} & = & \lambda_{i} v_{i}
\end{align*}
There's one problem left to solve: The rank of :math:`S_{W}` is at most :math:`(N-c)`, with :math:`N` samples and :math:`c` classes. In pattern recognition problems the number of samples :math:`N` is almost always samller than the dimension of the input data (the number of pixels), so the scatter matrix :math:`S_{W}` becomes singular (see [RJ91]_). In [BHK97]_ this was solved by performing a Principal Component Analysis on the data and projecting the samples into the :math:`(N-c)`-dimensional space. A Linear Discriminant Analysis was then performed on the reduced data, because :math:`S_{W}` isn't singular anymore.
The optimization problem can then be rewritten as:
.. math::
:nowrap:
\begin{align*}
W_{pca} & = & \operatorname{arg\,max}_{W} |W^T S_T W| \\
W_{fld} & = & \operatorname{arg\,max}_{W} \frac{|W^T W_{pca}^T S_{B} W_{pca} W|}{|W^T W_{pca}^T S_{W} W_{pca} W|}
\end{align*}
The transformation matrix :math:`W`, that projects a sample into the :math:`(c-1)`-dimensional space is then given by:
.. math::
W = W_{fld}^{T} W_{pca}^{T}
Fisherfaces in OpenCV
---------------------
.. literalinclude:: src/facerec_fisherfaces.cpp
:language: cpp
:linenos:
The source code for this demo application is also available in the ``src`` folder coming with this documentation:
* :download:`src/facerec_fisherfaces.cpp <src/facerec_fisherfaces.cpp>`
For this example I am going to use the Yale Facedatabase A, just because the plots are nicer. Each Fisherface has the same length as an original image, thus it can be displayed as an image. The demo shows (or saves) the first, at most 16 Fisherfaces:
.. image:: img/fisherfaces_opencv.png
:align: center
The Fisherfaces method learns a class-specific transformation matrix, so the they do not capture illumination as obviously as the Eigenfaces method. The Discriminant Analysis instead finds the facial features to discriminate between the persons. It's important to mention, that the performance of the Fisherfaces heavily depends on the input data as well. Practically said: if you learn the Fisherfaces for well-illuminated pictures only and you try to recognize faces in bad-illuminated scenes, then method is likely to find the wrong components (just because those features may not be predominant on bad illuminated images). This is somewhat logical, since the method had no chance to learn the illumination.
The Fisherfaces allow a reconstruction of the projected image, just like the Eigenfaces did. But since we only identified the features to distinguish between subjects, you can't expect a nice reconstruction of the original image. For the Fisherfaces method we'll project the sample image onto each of the Fisherfaces instead. So you'll have a nice visualization, which feature each of the Fisherfaces describes:
.. code-block:: cpp
// Display or save the image reconstruction at some predefined steps:
for(int num_component = 0; num_component < min(16, W.cols); num_component++) {
// Slice the Fisherface from the model:
Mat ev = W.col(num_component);
Mat projection = subspaceProject(ev, mean, images[0].reshape(1,1));
Mat reconstruction = subspaceReconstruct(ev, mean, projection);
// Normalize the result:
reconstruction = norm_0_255(reconstruction.reshape(1, images[0].rows));
// Display or save:
if(argc == 2) {
imshow(format("fisherface_reconstruction_%d", num_component), reconstruction);
} else {
imwrite(format("%s/fisherface_reconstruction_%d.png", output_folder.c_str(), num_component), reconstruction);
}
}
The differences may be subtle for the human eyes, but you should be able to see some differences:
.. image:: img/fisherface_reconstruction_opencv.png
:align: center
Local Binary Patterns Histograms
================================
Eigenfaces and Fisherfaces take a somewhat holistic approach to face recognition. You treat your data as a vector somewhere in a high-dimensional image space. We all know high-dimensionality is bad, so a lower-dimensional subspace is identified, where (probably) useful information is preserved. The Eigenfaces approach maximizes the total scatter, which can lead to problems if the variance is generated by an external source, because components with a maximum variance over all classes aren't necessarily useful for classification (see `http://www.bytefish.de/wiki/pca_lda_with_gnu_octave <http://www.bytefish.de/wiki/pca_lda_with_gnu_octave>`_). So to preserve some discriminative information we applied a Linear Discriminant Analysis and optimized as described in the Fisherfaces method. The Fisherfaces method worked great... at least for the constrained scenario we've assumed in our model.
Now real life isn't perfect. You simply can't guarantee perfect light settings in your images or 10 different images of a person. So what if there's only one image for each person? Our covariance estimates for the subspace *may* be horribly wrong, so will the recognition. Remember the Eigenfaces method had a 96% recognition rate on the AT&T Facedatabase? How many images do we actually need to get such useful estimates? Here are the Rank-1 recognition rates of the Eigenfaces and Fisherfaces method on the AT&T Facedatabase, which is a fairly easy image database:
.. image:: img/at_database_small_sample_size.png
:scale: 60%
:align: center
So in order to get good recognition rates you'll need at least 8(+-1) images for each person and the Fisherfaces method doesn't really help here. The above experiment is a 10-fold cross validated result carried out with the facerec framework at: `https://github.com/bytefish/facerec <https://github.com/bytefish/facerec>`_. This is not a publication, so I won't back these figures with a deep mathematical analysis. Please have a look into [KM01]_ for a detailed analysis of both methods, when it comes to small training datasets.
So some research concentrated on extracting local features from images. The idea is to not look at the whole image as a high-dimensional vector, but describe only local features of an object. The features you extract this way will have a low-dimensionality implicitly. A fine idea! But you'll soon observe the image representation we are given doesn't only suffer from illumination variations. Think of things like scale, translation or rotation in images - your local description has to be at least a bit robust against those things. Just like :ocv:class:`SIFT`, the Local Binary Patterns methodology has its roots in 2D texture analysis. The basic idea of Local Binary Patterns is to summarize the local structure in an image by comparing each pixel with its neighborhood. Take a pixel as center and threshold its neighbors against. If the intensity of the center pixel is greater-equal its neighbor, then denote it with 1 and 0 if not. You'll end up with a binary number for each pixel, just like 11001111. So with 8 surrounding pixels you'll end up with 2^8 possible combinations, called *Local Binary Patterns* or sometimes referred to as *LBP codes*. The first LBP operator described in literature actually used a fixed 3 x 3 neighborhood just like this:
.. image:: img/lbp/lbp.png
:scale: 80%
:align: center
Algorithmic Description
-----------------------
A more formal description of the LBP operator can be given as:
.. math::
LBP(x_c, y_c) = \sum_{p=0}^{P-1} 2^p s(i_p - i_c)
, with :math:`(x_c, y_c)` as central pixel with intensity :math:`i_c`; and :math:`i_n` being the intensity of the the neighbor pixel. :math:`s` is the sign function defined as:
.. math::
:nowrap:
\begin{equation}
s(x) =
\begin{cases}
1 & \text{if $x \geq 0$}\\
0 & \text{else}
\end{cases}
\end{equation}
This description enables you to capture very fine grained details in images. In fact the authors were able to compete with state of the art results for texture classification. Soon after the operator was published it was noted, that a fixed neighborhood fails to encode details differing in scale. So the operator was extended to use a variable neighborhood in [AHP04]_. The idea is to align an abritrary number of neighbors on a circle with a variable radius, which enables to capture the following neighborhoods:
.. image:: img/lbp/patterns.png
:scale: 80%
:align: center
For a given Point :math:`(x_c,y_c)` the position of the neighbor :math:`(x_p,y_p), p \in P` can be calculated by:
.. math::
:nowrap:
\begin{align*}
x_{p} & = & x_c + R \cos({\frac{2\pi p}{P}})\\
y_{p} & = & y_c - R \sin({\frac{2\pi p}{P}})
\end{align*}
Where :math:`R` is the radius of the circle and :math:`P` is the number of sample points.
The operator is an extension to the original LBP codes, so it's sometimes called *Extended LBP* (also referred to as *Circular LBP*) . If a points coordinate on the circle doesn't correspond to image coordinates, the point get's interpolated. Computer science has a bunch of clever interpolation schemes, the OpenCV implementation does a bilinear interpolation:
.. math::
:nowrap:
\begin{align*}
f(x,y) \approx \begin{bmatrix}
1-x & x \end{bmatrix} \begin{bmatrix}
f(0,0) & f(0,1) \\
f(1,0) & f(1,1) \end{bmatrix} \begin{bmatrix}
1-y \\
y \end{bmatrix}.
\end{align*}
By definition the LBP operator is robust against monotonic gray scale transformations. We can easily verify this by looking at the LBP image of an artificially modified image (so you see what an LBP image looks like!):
.. image:: img/lbp/lbp_yale.jpg
:scale: 60%
:align: center
So what's left to do is how to incorporate the spatial information in the face recognition model. The representation proposed by Ahonen et. al [AHP04]_ is to divide the LBP image into :math:`m` local regions and extract a histogram from each. The spatially enhanced feature vector is then obtained by concatenating the local histograms (**not merging them**). These histograms are called *Local Binary Patterns Histograms*.
Local Binary Patterns Histograms in OpenCV
------------------------------------------
.. literalinclude:: src/facerec_lbph.cpp
:language: cpp
:linenos:
The source code for this demo application is also available in the ``src`` folder coming with this documentation:
* :download:`src/facerec_lbph.cpp <src/facerec_lbph.cpp>`
Conclusion
==========
You've learned how to use the new :ocv:class:`FaceRecognizer` in real applications. After reading the document you also know how the algorithms work, so now it's time for you to experiment with the available algorithms. Use them, improve them and let the OpenCV community participate!
Credits
=======
This document wouldn't be possible without the kind permission to use the face images of the *AT&T Database of Faces* and the *Yale Facedatabase A/B*.
The Database of Faces
---------------------
** Important: when using these images, please give credit to "AT&T Laboratories, Cambridge." **
The Database of Faces, formerly *The ORL Database of Faces*, contains a set of face images taken between April 1992 and April 1994. The database was used in the context of a face recognition project carried out in collaboration with the Speech, Vision and Robotics Group of the Cambridge University Engineering Department.
There are ten different images of each of 40 distinct subjects. For some subjects, the images were taken at different times, varying the lighting, facial expressions (open / closed eyes, smiling / not smiling) and facial details (glasses / no glasses). All the images were taken against a dark homogeneous background with the subjects in an upright, frontal position (with tolerance for some side movement).
The files are in PGM format. The size of each image is 92x112 pixels, with 256 grey levels per pixel. The images are organised in 40 directories (one for each subject), which have names of the form sX, where X indicates the subject number (between 1 and 40). In each of these directories, there are ten different images of that subject, which have names of the form Y.pgm, where Y is the image number for that subject (between 1 and 10).
A copy of the database can be retrieved from: `http://www.cl.cam.ac.uk/research/dtg/attarchive/pub/data/att_faces.zip <http://www.cl.cam.ac.uk/research/dtg/attarchive/pub/data/att_faces.zip>`_.
Yale Facedatabase A
-------------------
*With the permission of the authors I am allowed to show a small number of images (say subject 1 and all the variations) and all images such as Fisherfaces and Eigenfaces from either Yale Facedatabase A or the Yale Facedatabase B.*
The Yale Face Database A (size 6.4MB) contains 165 grayscale images in GIF format of 15 individuals. There are 11 images per subject, one per different facial expression or configuration: center-light, w/glasses, happy, left-light, w/no glasses, normal, right-light, sad, sleepy, surprised, and wink. (Source: `http://cvc.yale.edu/projects/yalefaces/yalefaces.html <http://cvc.yale.edu/projects/yalefaces/yalefaces.html>`_)
Yale Facedatabase B
--------------------
*With the permission of the authors I am allowed to show a small number of images (say subject 1 and all the variations) and all images such as Fisherfaces and Eigenfaces from either Yale Facedatabase A or the Yale Facedatabase B.*
The extended Yale Face Database B contains 16128 images of 28 human subjects under 9 poses and 64 illumination conditions. The data format of this database is the same as the Yale Face Database B. Please refer to the homepage of the Yale Face Database B (or one copy of this page) for more detailed information of the data format.
You are free to use the extended Yale Face Database B for research purposes. All publications which use this database should acknowledge the use of "the Exteded Yale Face Database B" and reference Athinodoros Georghiades, Peter Belhumeur, and David Kriegman's paper, "From Few to Many: Illumination Cone Models for Face Recognition under Variable Lighting and Pose", PAMI, 2001, `[bibtex] <http://vision.ucsd.edu/~leekc/ExtYaleDatabase/athosref.html>`_.
The extended database as opposed to the original Yale Face Database B with 10 subjects was first reported by Kuang-Chih Lee, Jeffrey Ho, and David Kriegman in "Acquiring Linear Subspaces for Face Recognition under Variable Lighting, PAMI, May, 2005 `[pdf] <http://vision.ucsd.edu/~leekc/papers/9pltsIEEE.pdf>`_." All test image data used in the experiments are manually aligned, cropped, and then re-sized to 168x192 images. If you publish your experimental results with the cropped images, please reference the PAMI2005 paper as well. (Source: `http://vision.ucsd.edu/~leekc/ExtYaleDatabase/ExtYaleB.html <http://vision.ucsd.edu/~leekc/ExtYaleDatabase/ExtYaleB.html>`_)
Literature
==========
.. [AHP04] Ahonen, T., Hadid, A., and Pietikainen, M. *Face Recognition with Local Binary Patterns.* Computer Vision - ECCV 2004 (2004), 469–481.
.. [BHK97] Belhumeur, P. N., Hespanha, J., and Kriegman, D. *Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection.* IEEE Transactions on Pattern Analysis and Machine Intelligence 19, 7 (1997), 711–720.
.. [Bru92] Brunelli, R., Poggio, T. *Face Recognition through Geometrical Features.* European Conference on Computer Vision (ECCV) 1992, S. 792–800.
.. [Duda01] Duda, Richard O. and Hart, Peter E. and Stork, David G., *Pattern Classification* (2nd Edition) 2001.
.. [Fisher36] Fisher, R. A. *The use of multiple measurements in taxonomic problems.* Annals Eugen. 7 (1936), 179–188.
.. [GBK01] Georghiades, A.S. and Belhumeur, P.N. and Kriegman, D.J., *From Few to Many: Illumination Cone Models for Face Recognition under Variable Lighting and Pose* IEEE Transactions on Pattern Analysis and Machine Intelligence 23, 6 (2001), 643-660.
.. [Kanade73] Kanade, T. *Picture processing system by computer complex and recognition of human faces.* PhD thesis, Kyoto University, November 1973
.. [KM01] Martinez, A and Kak, A. *PCA versus LDA* IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 23, No.2, pp. 228-233, 2001.
.. [Lee05] Lee, K., Ho, J., Kriegman, D. *Acquiring Linear Subspaces for Face Recognition under Variable Lighting.* In: IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI) 27 (2005), Nr. 5
.. [Messer06] Messer, K. et al. *Performance Characterisation of Face Recognition Algorithms and Their Sensitivity to Severe Illumination Changes.* In: In: ICB, 2006, S. 1–11.
.. [RJ91] S. Raudys and A.K. Jain. *Small sample size effects in statistical pattern recognition: Recommendations for practitioneers.* - IEEE Transactions on Pattern Analysis and Machine Intelligence 13, 3 (1991), 252-264.
.. [Tan10] Tan, X., and Triggs, B. *Enhanced local texture feature sets for face recognition under difficult lighting conditions.* IEEE Transactions on Image Processing 19 (2010), 1635–650.
.. [TP91] Turk, M., and Pentland, A. *Eigenfaces for recognition.* Journal of Cognitive Neuroscience 3 (1991), 71–86.
.. [Tu06] Chiara Turati, Viola Macchi Cassia, F. S., and Leo, I. *Newborns face recognition: Role of inner and outer facial features. Child Development* 77, 2 (2006), 297–311.
.. [Wiskott97] Wiskott, L., Fellous, J., Krüger, N., Malsburg, C. *Face Recognition By Elastic Bunch Graph Matching.* IEEE Transactions on Pattern Analysis and Machine Intelligence 19 (1997), S. 775–779
.. [Zhao03] Zhao, W., Chellappa, R., Phillips, P., and Rosenfeld, A. Face recognition: A literature survey. ACM Computing Surveys (CSUR) 35, 4 (2003), 399–458.
.. _appendixft:
Appendix
========
Creating the CSV File
---------------------
You don't really want to create the CSV file by hand. I have prepared you a little Python script ``create_csv.py`` (you find it at ``/src/create_csv.py`` coming with this tutorial) that automatically creates you a CSV file. If you have your images in hierarchie like this (``/basepath/<subject>/<image.ext>``):
.. code-block:: none
philipp@mango:~/facerec/data/at$ tree
.
|-- s1
| |-- 1.pgm
| |-- ...
| |-- 10.pgm
|-- s2
| |-- 1.pgm
| |-- ...
| |-- 10.pgm
...
|-- s40
| |-- 1.pgm
| |-- ...
| |-- 10.pgm
Then simply call ``create_csv.py`` with the path to the folder, just like this and you could save the output:
.. code-block:: none
philipp@mango:~/facerec/data$ python create_csv.py
at/s13/2.pgm;0
at/s13/7.pgm;0
at/s13/6.pgm;0
at/s13/9.pgm;0
at/s13/5.pgm;0
at/s13/3.pgm;0
at/s13/4.pgm;0
at/s13/10.pgm;0
at/s13/8.pgm;0
at/s13/1.pgm;0
at/s17/2.pgm;1
at/s17/7.pgm;1
at/s17/6.pgm;1
at/s17/9.pgm;1
at/s17/5.pgm;1
at/s17/3.pgm;1
[...]
Here is the script, if you can't find it:
.. literalinclude:: ./src/create_csv.py
:language: python
:linenos:
Aligning Face Images
---------------------
An accurate alignment of your image data is especially important in tasks like emotion detection, were you need as much detail as possible. Believe me... You don't want to do this by hand. So I've prepared you a tiny Python script. The code is really easy to use. To scale, rotate and crop the face image you just need to call *CropFace(image, eye_left, eye_right, offset_pct, dest_sz)*, where:
* *eye_left* is the position of the left eye
* *eye_right* is the position of the right eye
* *offset_pct* is the percent of the image you want to keep next to the eyes (horizontal, vertical direction)
* *dest_sz* is the size of the output image
If you are using the same *offset_pct* and *dest_sz* for your images, they are all aligned at the eyes.
.. literalinclude:: ./src/crop_face.py
:language: python
:linenos:
Imagine we are given `this photo of Arnold Schwarzenegger <http://en.wikipedia.org/wiki/File:Arnold_Schwarzenegger_edit%28ws%29.jpg>`_, which is under a Public Domain license. The (x,y)-position of the eyes is approximately *(252,364)* for the left and *(420,366)* for the right eye. Now you only need to define the horizontal offset, vertical offset and the size your scaled, rotated & cropped face should have.
Here are some examples:
+---------------------------------+----------------------------------------------------------------------------+
| Configuration | Cropped, Scaled, Rotated Face |
+=================================+============================================================================+
| 0.1 (10%), 0.1 (10%), (200,200) | .. image:: ./img/tutorial/gender_classification/arnie_10_10_200_200.jpg |
+---------------------------------+----------------------------------------------------------------------------+
| 0.2 (20%), 0.2 (20%), (200,200) | .. image:: ./img/tutorial/gender_classification/arnie_20_20_200_200.jpg |
+---------------------------------+----------------------------------------------------------------------------+
| 0.3 (30%), 0.3 (30%), (200,200) | .. image:: ./img/tutorial/gender_classification/arnie_30_30_200_200.jpg |
+---------------------------------+----------------------------------------------------------------------------+
| 0.2 (20%), 0.2 (20%), (70,70) | .. image:: ./img/tutorial/gender_classification/arnie_20_20_70_70.jpg |
+---------------------------------+----------------------------------------------------------------------------+
CSV for the AT&T Facedatabase
------------------------------
.. literalinclude:: etc/at.txt
:language: none
:linenos:

BIN
modules/contrib/doc/facerec/img/at_database_small_sample_size.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 33 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_autumn.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.3 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_bone.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.3 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_cool.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.3 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_hot.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.4 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_hsv.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.7 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_jet.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.5 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_mkpj1.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.5 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_mkpj2.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.5 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_ocean.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.4 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_pink.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.4 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_rainbow.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.5 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_spring.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.2 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_summer.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.3 KiB

BIN
modules/contrib/doc/facerec/img/colormaps/colorscale_winter.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.2 KiB

BIN
modules/contrib/doc/facerec/img/eigenface_reconstruction_opencv.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 171 KiB

BIN
modules/contrib/doc/facerec/img/eigenfaces_opencv.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 108 KiB

BIN
modules/contrib/doc/facerec/img/fisherface_reconstruction_opencv.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 111 KiB

BIN
modules/contrib/doc/facerec/img/fisherfaces_opencv.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 281 KiB

BIN
modules/contrib/doc/facerec/img/lbp/lbp.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 15 KiB

BIN
modules/contrib/doc/facerec/img/lbp/lbp_yale.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 83 KiB

BIN
modules/contrib/doc/facerec/img/lbp/patterns.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 18 KiB

BIN
modules/contrib/doc/facerec/img/tutorial/facerec_video/facerec_video.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 290 KiB

BIN
modules/contrib/doc/facerec/img/tutorial/gender_classification/arnie_10_10_200_200.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 5.4 KiB

BIN
modules/contrib/doc/facerec/img/tutorial/gender_classification/arnie_20_20_200_200.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 6.2 KiB

BIN
modules/contrib/doc/facerec/img/tutorial/gender_classification/arnie_20_20_70_70.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.8 KiB

BIN
modules/contrib/doc/facerec/img/tutorial/gender_classification/arnie_30_30_200_200.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 7.1 KiB

BIN
modules/contrib/doc/facerec/img/tutorial/gender_classification/clooney_set.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 92 KiB

BIN
modules/contrib/doc/facerec/img/tutorial/gender_classification/fisherface_0.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 36 KiB

BIN
modules/contrib/doc/facerec/img/tutorial/gender_classification/fisherface_reconstruction_0.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 10 KiB

BIN
modules/contrib/doc/facerec/img/tutorial/gender_classification/mean.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 9.9 KiB

32
modules/contrib/doc/facerec/index.rst
View File

@@ -1,32 +0,0 @@
FaceRecognizer - Face Recognition with OpenCV
##############################################
OpenCV 2.4 now comes with the very new :ocv:class:`FaceRecognizer` class for face recognition. This documentation is going to explain you :doc:`the API <facerec_api>` in detail and it will give you a lot of help to get started (full source code examples). :doc:`Face Recognition with OpenCV <facerec_tutorial>` is the definite guide to the new :ocv:class:`FaceRecognizer`. There's also a :doc:`tutorial on gender classification <tutorial/facerec_gender_classification>`, a :doc:`tutorial for face recognition in videos <tutorial/facerec_video_recognition>` and it's shown :doc:`how to load & save your results <tutorial/facerec_save_load>`.
These documents are the help I have wished for, when I was working myself into face recognition. I hope you also think the new :ocv:class:`FaceRecognizer` is a useful addition to OpenCV.
Please issue any feature requests and/or bugs on the official OpenCV bug tracker at:
* http://code.opencv.org/projects/opencv/issues
Contents
========
.. toctree::
:maxdepth: 1
FaceRecognizer API <facerec_api>
Guide to Face Recognition with OpenCV <facerec_tutorial>
Tutorial on Gender Classification <tutorial/facerec_gender_classification>
Tutorial on Face Recognition in Videos <tutorial/facerec_video_recognition>
Tutorial On Saving & Loading a FaceRecognizer <tutorial/facerec_save_load>
How to use Colormaps in OpenCV <colormaps>
Changelog <facerec_changelog>
Indices and tables
==================
* :ref:`genindex`
* :ref:`modindex`
* :ref:`search`

25
modules/contrib/doc/facerec/src/CMakeLists.txt
View File

@@ -1,25 +0,0 @@
CMAKE_MINIMUM_REQUIRED(VERSION 2.6)
set(name "facerec")
project(facerec_cpp_samples)
#SET(OpenCV_DIR /path/to/your/opencv/installation)
# packages
find_package(OpenCV REQUIRED) # http://opencv.org
# probably you should loop through the sample files here
add_executable(facerec_demo facerec_demo.cpp)
target_link_libraries(facerec_demo opencv_core opencv_contrib opencv_imgproc opencv_highgui)
add_executable(facerec_video facerec_video.cpp)
target_link_libraries(facerec_video opencv_contrib opencv_core opencv_imgproc opencv_highgui opencv_objdetect opencv_imgproc)
add_executable(facerec_eigenfaces facerec_eigenfaces.cpp)
target_link_libraries(facerec_eigenfaces opencv_contrib opencv_core opencv_imgproc opencv_highgui)
add_executable(facerec_fisherfaces facerec_fisherfaces.cpp)
target_link_libraries(facerec_fisherfaces opencv_contrib opencv_core opencv_imgproc opencv_highgui)
add_executable(facerec_lbph facerec_lbph.cpp)
target_link_libraries(facerec_lbph opencv_contrib opencv_core opencv_imgproc opencv_highgui)

43
modules/contrib/doc/facerec/src/create_csv.py
View File

@@ -1,43 +0,0 @@
#!/usr/bin/env python
import sys
import os.path
# This is a tiny script to help you creating a CSV file from a face
# database with a similar hierarchie:
#
# philipp@mango:~/facerec/data/at$ tree
# .
# |-- README
# |-- s1
# | |-- 1.pgm
# | |-- ...
# | |-- 10.pgm
# |-- s2
# | |-- 1.pgm
# | |-- ...
# | |-- 10.pgm
# ...
# |-- s40
# | |-- 1.pgm
# | |-- ...
# | |-- 10.pgm
#
if __name__ == "__main__":
if len(sys.argv) != 2:
print "usage: create_csv <base_path>"
sys.exit(1)
BASE_PATH=sys.argv[1]
SEPARATOR=";"
label = 0
for dirname, dirnames, filenames in os.walk(BASE_PATH):
for subdirname in dirnames:
subject_path = os.path.join(dirname, subdirname)
for filename in os.listdir(subject_path):
abs_path = "%s/%s" % (subject_path, filename)
print "%s%s%d" % (abs_path, SEPARATOR, label)
label = label + 1

112
modules/contrib/doc/facerec/src/crop_face.py
View File

@@ -1,112 +0,0 @@
#!/usr/bin/env python
# Software License Agreement (BSD License)
#
# Copyright (c) 2012, Philipp Wagner
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following
# disclaimer in the documentation and/or other materials provided
# with the distribution.
# * Neither the name of the author nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
# FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
# COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
# BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
# ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
# POSSIBILITY OF SUCH DAMAGE.
import sys, math, Image
def Distance(p1,p2):
dx = p2[0] - p1[0]
dy = p2[1] - p1[1]
return math.sqrt(dx*dx+dy*dy)
def ScaleRotateTranslate(image, angle, center = None, new_center = None, scale = None, resample=Image.BICUBIC):
if (scale is None) and (center is None):
return image.rotate(angle=angle, resample=resample)
nx,ny = x,y = center
sx=sy=1.0
if new_center:
(nx,ny) = new_center
if scale:
(sx,sy) = (scale, scale)
cosine = math.cos(angle)
sine = math.sin(angle)
a = cosine/sx
b = sine/sx
c = x-nx*a-ny*b
d = -sine/sy
e = cosine/sy
f = y-nx*d-ny*e
return image.transform(image.size, Image.AFFINE, (a,b,c,d,e,f), resample=resample)
def CropFace(image, eye_left=(0,0), eye_right=(0,0), offset_pct=(0.2,0.2), dest_sz = (70,70)):
# calculate offsets in original image
offset_h = math.floor(float(offset_pct[0])*dest_sz[0])
offset_v = math.floor(float(offset_pct[1])*dest_sz[1])
# get the direction
eye_direction = (eye_right[0] - eye_left[0], eye_right[1] - eye_left[1])
# calc rotation angle in radians
rotation = -math.atan2(float(eye_direction[1]),float(eye_direction[0]))
# distance between them
dist = Distance(eye_left, eye_right)
# calculate the reference eye-width
reference = dest_sz[0] - 2.0*offset_h
# scale factor
scale = float(dist)/float(reference)
# rotate original around the left eye
image = ScaleRotateTranslate(image, center=eye_left, angle=rotation)
# crop the rotated image
crop_xy = (eye_left[0] - scale*offset_h, eye_left[1] - scale*offset_v)
crop_size = (dest_sz[0]*scale, dest_sz[1]*scale)
image = image.crop((int(crop_xy[0]), int(crop_xy[1]), int(crop_xy[0]+crop_size[0]), int(crop_xy[1]+crop_size[1])))
# resize it
image = image.resize(dest_sz, Image.ANTIALIAS)
return image
def readFileNames():
try:
inFile = open('path_to_created_csv_file.csv')
except:
raise IOError('There is no file named path_to_created_csv_file.csv in current directory.')
return False
picPath = []
picIndex = []
for line in inFile.readlines():
if line != '':
fields = line.rstrip().split(';')
picPath.append(fields[0])
picIndex.append(int(fields[1]))
return (picPath, picIndex)
if __name__ == "__main__":
[images, indexes]=readFileNames()
if not os.path.exists("modified"):
os.makedirs("modified")
for img in images:
image = Image.open(img)
CropFace(image, eye_left=(252,364), eye_right=(420,366), offset_pct=(0.1,0.1), dest_sz=(200,200)).save("modified/"+img.rstrip().split('/')[1]+"_10_10_200_200.jpg")
CropFace(image, eye_left=(252,364), eye_right=(420,366), offset_pct=(0.2,0.2), dest_sz=(200,200)).save("modified/"+img.rstrip().split('/')[1]+"_20_20_200_200.jpg")
CropFace(image, eye_left=(252,364), eye_right=(420,366), offset_pct=(0.3,0.3), dest_sz=(200,200)).save("modified/"+img.rstrip().split('/')[1]+"_30_30_200_200.jpg")
CropFace(image, eye_left=(252,364), eye_right=(420,366), offset_pct=(0.2,0.2)).save("modified/"+img.rstrip().split('/')[1]+"_20_20_70_70.jpg")

169
modules/contrib/doc/facerec/src/facerec_demo.cpp
View File

@@ -1,169 +0,0 @@
/*
* Copyright (c) 2011. Philipp Wagner <bytefish[at]gmx[dot]de>.
* Released to public domain under terms of the BSD Simplified license.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of the organization nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* See <http://www.opensource.org/licenses/bsd-license>
*/
#include "opencv2/core.hpp"
#include "opencv2/contrib.hpp"
#include "opencv2/highgui.hpp"
#include <iostream>
#include <fstream>
#include <sstream>
using namespace cv;
using namespace std;
static Mat norm_0_255(InputArray _src) {
Mat src = _src.getMat();
// Create and return normalized image:
Mat dst;
switch(src.channels()) {
case 1:
cv::normalize(_src, dst, 0, 255, NORM_MINMAX, CV_8UC1);
break;
case 3:
cv::normalize(_src, dst, 0, 255, NORM_MINMAX, CV_8UC3);
break;
default:
src.copyTo(dst);
break;
}
return dst;
}
static void read_csv(const string& filename, vector<Mat>& images, vector<int>& labels, char separator = ';') {
std::ifstream file(filename.c_str(), ifstream::in);
if (!file) {
string error_message = "No valid input file was given, please check the given filename.";
CV_Error(CV_StsBadArg, error_message);
}
string line, path, classlabel;
while (getline(file, line)) {
stringstream liness(line);
getline(liness, path, separator);
getline(liness, classlabel);
if(!path.empty() && !classlabel.empty()) {
images.push_back(imread(path, 0));
labels.push_back(atoi(classlabel.c_str()));
}
}
}
int main(int argc, const char *argv[]) {
// Check for valid command line arguments, print usage
// if no arguments were given.
if (argc != 2) {
cout << "usage: " << argv[0] << " <csv.ext>" << endl;
exit(1);
}
// Get the path to your CSV.
string fn_csv = string(argv[1]);
// These vectors hold the images and corresponding labels.
vector<Mat> images;
vector<int> labels;
// Read in the data. This can fail if no valid
// input filename is given.
try {
read_csv(fn_csv, images, labels);
} catch (cv::Exception& e) {
cerr << "Error opening file \"" << fn_csv << "\". Reason: " << e.msg << endl;
// nothing more we can do
exit(1);
}
// Quit if there are not enough images for this demo.
if(images.size() <= 1) {
string error_message = "This demo needs at least 2 images to work. Please add more images to your data set!";
CV_Error(CV_StsError, error_message);
}
// Get the height from the first image. We'll need this
// later in code to reshape the images to their original
// size:
int height = images[0].rows;
// The following lines simply get the last images from
// your dataset and remove it from the vector. This is
// done, so that the training data (which we learn the
// cv::FaceRecognizer on) and the test data we test
// the model with, do not overlap.
Mat testSample = images[images.size() - 1];
int testLabel = labels[labels.size() - 1];
images.pop_back();
labels.pop_back();
// The following lines create an Eigenfaces model for
// face recognition and train it with the images and
// labels read from the given CSV file.
// This here is a full PCA, if you just want to keep
// 10 principal components (read Eigenfaces), then call
// the factory method like this:
//
// cv::createEigenFaceRecognizer(10);
//
// If you want to create a FaceRecognizer with a
// confidennce threshold, call it with:
//
// cv::createEigenFaceRecognizer(10, 123.0);
//
Ptr<FaceRecognizer> model = createFisherFaceRecognizer();
model->train(images, labels);
// The following line predicts the label of a given
// test image:
int predictedLabel = model->predict(testSample);
//
// To get the confidence of a prediction call the model with:
//
// int predictedLabel = -1;
// double confidence = 0.0;
// model->predict(testSample, predictedLabel, confidence);
//
string result_message = format("Predicted class = %d / Actual class = %d.", predictedLabel, testLabel);
cout << result_message << endl;
// Sometimes you'll need to get/set internal model data,
// which isn't exposed by the public cv::FaceRecognizer.
// Since each cv::FaceRecognizer is derived from a
// cv::Algorithm, you can query the data.
//
// First we'll use it to set the threshold of the FaceRecognizer
// to 0.0 without retraining the model. This can be useful if
// you are evaluating the model:
//
model->set("threshold", 0.0);
// Now the threshold of this model is set to 0.0. A prediction
// now returns -1, as it's impossible to have a distance below
// it
predictedLabel = model->predict(testSample);
cout << "Predicted class = " << predictedLabel << endl;
// Here is how to get the eigenvalues of this Eigenfaces model:
Mat eigenvalues = model->getMat("eigenvalues");
// And we can do the same to display the Eigenvectors (read Eigenfaces):
Mat W = model->getMat("eigenvectors");
// From this we will display the (at most) first 10 Eigenfaces:
for (int i = 0; i < min(10, W.cols); i++) {
string msg = format("Eigenvalue #%d = %.5f", i, eigenvalues.at<double>(i));
cout << msg << endl;
// get eigenvector #i
Mat ev = W.col(i).clone();
// Reshape to original size & normalize to [0...255] for imshow.
Mat grayscale = norm_0_255(ev.reshape(1, height));
// Show the image & apply a Jet colormap for better sensing.
Mat cgrayscale;
applyColorMap(grayscale, cgrayscale, COLORMAP_JET);
imshow(format("%d", i), cgrayscale);
}
waitKey(0);
return 0;
}

193
modules/contrib/doc/facerec/src/facerec_eigenfaces.cpp
View File

@@ -1,193 +0,0 @@
/*
* Copyright (c) 2011. Philipp Wagner <bytefish[at]gmx[dot]de>.
* Released to public domain under terms of the BSD Simplified license.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of the organization nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* See <http://www.opensource.org/licenses/bsd-license>
*/
#include "opencv2/core.hpp"
#include "opencv2/contrib.hpp"
#include "opencv2/highgui.hpp"
#include <iostream>
#include <fstream>
#include <sstream>
using namespace cv;
using namespace std;
static Mat norm_0_255(InputArray _src) {
Mat src = _src.getMat();
// Create and return normalized image:
Mat dst;
switch(src.channels()) {
case 1:
cv::normalize(_src, dst, 0, 255, NORM_MINMAX, CV_8UC1);
break;
case 3:
cv::normalize(_src, dst, 0, 255, NORM_MINMAX, CV_8UC3);
break;
default:
src.copyTo(dst);
break;
}
return dst;
}
static void read_csv(const string& filename, vector<Mat>& images, vector<int>& labels, char separator = ';') {
std::ifstream file(filename.c_str(), ifstream::in);
if (!file) {
string error_message = "No valid input file was given, please check the given filename.";
CV_Error(CV_StsBadArg, error_message);
}
string line, path, classlabel;
while (getline(file, line)) {
stringstream liness(line);
getline(liness, path, separator);
getline(liness, classlabel);
if(!path.empty() && !classlabel.empty()) {
images.push_back(imread(path, 0));
labels.push_back(atoi(classlabel.c_str()));
}
}
}
int main(int argc, const char *argv[]) {
// Check for valid command line arguments, print usage
// if no arguments were given.
if (argc < 2) {
cout << "usage: " << argv[0] << " <csv.ext> <output_folder> " << endl;
exit(1);
}
string output_folder = ".";
if (argc == 3) {
output_folder = string(argv[2]);
}
// Get the path to your CSV.
string fn_csv = string(argv[1]);
// These vectors hold the images and corresponding labels.
vector<Mat> images;
vector<int> labels;
// Read in the data. This can fail if no valid
// input filename is given.
try {
read_csv(fn_csv, images, labels);
} catch (cv::Exception& e) {
cerr << "Error opening file \"" << fn_csv << "\". Reason: " << e.msg << endl;
// nothing more we can do
exit(1);
}
// Quit if there are not enough images for this demo.
if(images.size() <= 1) {
string error_message = "This demo needs at least 2 images to work. Please add more images to your data set!";
CV_Error(CV_StsError, error_message);
}
// Get the height from the first image. We'll need this
// later in code to reshape the images to their original
// size:
int height = images[0].rows;
// The following lines simply get the last images from
// your dataset and remove it from the vector. This is
// done, so that the training data (which we learn the
// cv::FaceRecognizer on) and the test data we test
// the model with, do not overlap.
Mat testSample = images[images.size() - 1];
int testLabel = labels[labels.size() - 1];
images.pop_back();
labels.pop_back();
// The following lines create an Eigenfaces model for
// face recognition and train it with the images and
// labels read from the given CSV file.
// This here is a full PCA, if you just want to keep
// 10 principal components (read Eigenfaces), then call
// the factory method like this:
//
// cv::createEigenFaceRecognizer(10);
//
// If you want to create a FaceRecognizer with a
// confidence threshold (e.g. 123.0), call it with:
//
// cv::createEigenFaceRecognizer(10, 123.0);
//
// If you want to use _all_ Eigenfaces and have a threshold,
// then call the method like this:
//
// cv::createEigenFaceRecognizer(0, 123.0);
//
Ptr<FaceRecognizer> model = createEigenFaceRecognizer();
model->train(images, labels);
// The following line predicts the label of a given
// test image:
int predictedLabel = model->predict(testSample);
//
// To get the confidence of a prediction call the model with:
//
// int predictedLabel = -1;
// double confidence = 0.0;
// model->predict(testSample, predictedLabel, confidence);
//
string result_message = format("Predicted class = %d / Actual class = %d.", predictedLabel, testLabel);
cout << result_message << endl;
// Here is how to get the eigenvalues of this Eigenfaces model:
Mat eigenvalues = model->getMat("eigenvalues");
// And we can do the same to display the Eigenvectors (read Eigenfaces):
Mat W = model->getMat("eigenvectors");
// Get the sample mean from the training data
Mat mean = model->getMat("mean");
// Display or save:
if(argc == 2) {
imshow("mean", norm_0_255(mean.reshape(1, images[0].rows)));
} else {
imwrite(format("%s/mean.png", output_folder.c_str()), norm_0_255(mean.reshape(1, images[0].rows)));
}
// Display or save the Eigenfaces:
for (int i = 0; i < min(10, W.cols); i++) {
string msg = format("Eigenvalue #%d = %.5f", i, eigenvalues.at<double>(i));
cout << msg << endl;
// get eigenvector #i
Mat ev = W.col(i).clone();
// Reshape to original size & normalize to [0...255] for imshow.
Mat grayscale = norm_0_255(ev.reshape(1, height));
// Show the image & apply a Jet colormap for better sensing.
Mat cgrayscale;
applyColorMap(grayscale, cgrayscale, COLORMAP_JET);
// Display or save:
if(argc == 2) {
imshow(format("eigenface_%d", i), cgrayscale);
} else {
imwrite(format("%s/eigenface_%d.png", output_folder.c_str(), i), norm_0_255(cgrayscale));
}
}
// Display or save the image reconstruction at some predefined steps:
for(int num_components = min(W.cols, 10); num_components < min(W.cols, 300); num_components+=15) {
// slice the eigenvectors from the model
Mat evs = Mat(W, Range::all(), Range(0, num_components));
Mat projection = subspaceProject(evs, mean, images[0].reshape(1,1));
Mat reconstruction = subspaceReconstruct(evs, mean, projection);
// Normalize the result:
reconstruction = norm_0_255(reconstruction.reshape(1, images[0].rows));
// Display or save:
if(argc == 2) {
imshow(format("eigenface_reconstruction_%d", num_components), reconstruction);
} else {
imwrite(format("%s/eigenface_reconstruction_%d.png", output_folder.c_str(), num_components), reconstruction);
}
}
// Display if we are not writing to an output folder:
if(argc == 2) {
waitKey(0);
}
return 0;
}

191
modules/contrib/doc/facerec/src/facerec_fisherfaces.cpp
View File

@@ -1,191 +0,0 @@
/*
* Copyright (c) 2011. Philipp Wagner <bytefish[at]gmx[dot]de>.
* Released to public domain under terms of the BSD Simplified license.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of the organization nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* See <http://www.opensource.org/licenses/bsd-license>
*/
#include "opencv2/core.hpp"
#include "opencv2/contrib.hpp"
#include "opencv2/highgui.hpp"
#include <iostream>
#include <fstream>
#include <sstream>
using namespace cv;
using namespace std;
static Mat norm_0_255(InputArray _src) {
Mat src = _src.getMat();
// Create and return normalized image:
Mat dst;
switch(src.channels()) {
case 1:
cv::normalize(_src, dst, 0, 255, NORM_MINMAX, CV_8UC1);
break;
case 3:
cv::normalize(_src, dst, 0, 255, NORM_MINMAX, CV_8UC3);
break;
default:
src.copyTo(dst);
break;
}
return dst;
}
static void read_csv(const string& filename, vector<Mat>& images, vector<int>& labels, char separator = ';') {
std::ifstream file(filename.c_str(), ifstream::in);
if (!file) {
string error_message = "No valid input file was given, please check the given filename.";
CV_Error(CV_StsBadArg, error_message);
}
string line, path, classlabel;
while (getline(file, line)) {
stringstream liness(line);
getline(liness, path, separator);
getline(liness, classlabel);
if(!path.empty() && !classlabel.empty()) {
images.push_back(imread(path, 0));
labels.push_back(atoi(classlabel.c_str()));
}
}
}
int main(int argc, const char *argv[]) {
// Check for valid command line arguments, print usage
// if no arguments were given.
if (argc < 2) {
cout << "usage: " << argv[0] << " <csv.ext> <output_folder> " << endl;
exit(1);
}
string output_folder = ".";
if (argc == 3) {
output_folder = string(argv[2]);
}
// Get the path to your CSV.
string fn_csv = string(argv[1]);
// These vectors hold the images and corresponding labels.
vector<Mat> images;
vector<int> labels;
// Read in the data. This can fail if no valid
// input filename is given.
try {
read_csv(fn_csv, images, labels);
} catch (cv::Exception& e) {
cerr << "Error opening file \"" << fn_csv << "\". Reason: " << e.msg << endl;
// nothing more we can do
exit(1);
}
// Quit if there are not enough images for this demo.
if(images.size() <= 1) {
string error_message = "This demo needs at least 2 images to work. Please add more images to your data set!";
CV_Error(CV_StsError, error_message);
}
// Get the height from the first image. We'll need this
// later in code to reshape the images to their original
// size:
int height = images[0].rows;
// The following lines simply get the last images from
// your dataset and remove it from the vector. This is
// done, so that the training data (which we learn the
// cv::FaceRecognizer on) and the test data we test
// the model with, do not overlap.
Mat testSample = images[images.size() - 1];
int testLabel = labels[labels.size() - 1];
images.pop_back();
labels.pop_back();
// The following lines create an Fisherfaces model for
// face recognition and train it with the images and
// labels read from the given CSV file.
// If you just want to keep 10 Fisherfaces, then call
// the factory method like this:
//
// cv::createFisherFaceRecognizer(10);
//
// However it is not useful to discard Fisherfaces! Please
// always try to use _all_ available Fisherfaces for
// classification.
//
// If you want to create a FaceRecognizer with a
// confidence threshold (e.g. 123.0) and use _all_
// Fisherfaces, then call it with:
//
// cv::createFisherFaceRecognizer(0, 123.0);
//
Ptr<FaceRecognizer> model = createFisherFaceRecognizer();
model->train(images, labels);
// The following line predicts the label of a given
// test image:
int predictedLabel = model->predict(testSample);
//
// To get the confidence of a prediction call the model with:
//
// int predictedLabel = -1;
// double confidence = 0.0;
// model->predict(testSample, predictedLabel, confidence);
//
string result_message = format("Predicted class = %d / Actual class = %d.", predictedLabel, testLabel);
cout << result_message << endl;
// Here is how to get the eigenvalues of this Eigenfaces model:
Mat eigenvalues = model->getMat("eigenvalues");
// And we can do the same to display the Eigenvectors (read Eigenfaces):
Mat W = model->getMat("eigenvectors");
// Get the sample mean from the training data
Mat mean = model->getMat("mean");
// Display or save:
if(argc == 2) {
imshow("mean", norm_0_255(mean.reshape(1, images[0].rows)));
} else {
imwrite(format("%s/mean.png", output_folder.c_str()), norm_0_255(mean.reshape(1, images[0].rows)));
}
// Display or save the first, at most 16 Fisherfaces:
for (int i = 0; i < min(16, W.cols); i++) {
string msg = format("Eigenvalue #%d = %.5f", i, eigenvalues.at<double>(i));
cout << msg << endl;
// get eigenvector #i
Mat ev = W.col(i).clone();
// Reshape to original size & normalize to [0...255] for imshow.
Mat grayscale = norm_0_255(ev.reshape(1, height));
// Show the image & apply a Bone colormap for better sensing.
Mat cgrayscale;
applyColorMap(grayscale, cgrayscale, COLORMAP_BONE);
// Display or save:
if(argc == 2) {
imshow(format("fisherface_%d", i), cgrayscale);
} else {
imwrite(format("%s/fisherface_%d.png", output_folder.c_str(), i), norm_0_255(cgrayscale));
}
}
// Display or save the image reconstruction at some predefined steps:
for(int num_component = 0; num_component < min(16, W.cols); num_component++) {
// Slice the Fisherface from the model:
Mat ev = W.col(num_component);
Mat projection = subspaceProject(ev, mean, images[0].reshape(1,1));
Mat reconstruction = subspaceReconstruct(ev, mean, projection);
// Normalize the result:
reconstruction = norm_0_255(reconstruction.reshape(1, images[0].rows));
// Display or save:
if(argc == 2) {
imshow(format("fisherface_reconstruction_%d", num_component), reconstruction);
} else {
imwrite(format("%s/fisherface_reconstruction_%d.png", output_folder.c_str(), num_component), reconstruction);
}
}
// Display if we are not writing to an output folder:
if(argc == 2) {
waitKey(0);
}
return 0;
}

155
modules/contrib/doc/facerec/src/facerec_lbph.cpp
View File

@@ -1,155 +0,0 @@
/*
* Copyright (c) 2011. Philipp Wagner <bytefish[at]gmx[dot]de>.
* Released to public domain under terms of the BSD Simplified license.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of the organization nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* See <http://www.opensource.org/licenses/bsd-license>
*/
#include "opencv2/core.hpp"
#include "opencv2/contrib.hpp"
#include "opencv2/highgui.hpp"
#include <iostream>
#include <fstream>
#include <sstream>
using namespace cv;
using namespace std;
static void read_csv(const string& filename, vector<Mat>& images, vector<int>& labels, char separator = ';') {
std::ifstream file(filename.c_str(), ifstream::in);
if (!file) {
string error_message = "No valid input file was given, please check the given filename.";
CV_Error(CV_StsBadArg, error_message);
}
string line, path, classlabel;
while (getline(file, line)) {
stringstream liness(line);
getline(liness, path, separator);
getline(liness, classlabel);
if(!path.empty() && !classlabel.empty()) {
images.push_back(imread(path, 0));
labels.push_back(atoi(classlabel.c_str()));
}
}
}
int main(int argc, const char *argv[]) {
// Check for valid command line arguments, print usage
// if no arguments were given.
if (argc != 2) {
cout << "usage: " << argv[0] << " <csv.ext>" << endl;
exit(1);
}
// Get the path to your CSV.
string fn_csv = string(argv[1]);
// These vectors hold the images and corresponding labels.
vector<Mat> images;
vector<int> labels;
// Read in the data. This can fail if no valid
// input filename is given.
try {
read_csv(fn_csv, images, labels);
} catch (cv::Exception& e) {
cerr << "Error opening file \"" << fn_csv << "\". Reason: " << e.msg << endl;
// nothing more we can do
exit(1);
}
// Quit if there are not enough images for this demo.
if(images.size() <= 1) {
string error_message = "This demo needs at least 2 images to work. Please add more images to your data set!";
CV_Error(CV_StsError, error_message);
}
// Get the height from the first image. We'll need this
// later in code to reshape the images to their original
// size:
int height = images[0].rows;
// The following lines simply get the last images from
// your dataset and remove it from the vector. This is
// done, so that the training data (which we learn the
// cv::FaceRecognizer on) and the test data we test
// the model with, do not overlap.
Mat testSample = images[images.size() - 1];
int testLabel = labels[labels.size() - 1];
images.pop_back();
labels.pop_back();
// The following lines create an LBPH model for
// face recognition and train it with the images and
// labels read from the given CSV file.
//
// The LBPHFaceRecognizer uses Extended Local Binary Patterns
// (it's probably configurable with other operators at a later
// point), and has the following default values
//
// radius = 1
// neighbors = 8
// grid_x = 8
// grid_y = 8
//
// So if you want a LBPH FaceRecognizer using a radius of
// 2 and 16 neighbors, call the factory method with:
//
// cv::createLBPHFaceRecognizer(2, 16);
//
// And if you want a threshold (e.g. 123.0) call it with its default values:
//
// cv::createLBPHFaceRecognizer(1,8,8,8,123.0)
//
Ptr<FaceRecognizer> model = createLBPHFaceRecognizer();
model->train(images, labels);
// The following line predicts the label of a given
// test image:
int predictedLabel = model->predict(testSample);
//
// To get the confidence of a prediction call the model with:
//
// int predictedLabel = -1;
// double confidence = 0.0;
// model->predict(testSample, predictedLabel, confidence);
//
string result_message = format("Predicted class = %d / Actual class = %d.", predictedLabel, testLabel);
cout << result_message << endl;
// Sometimes you'll need to get/set internal model data,
// which isn't exposed by the public cv::FaceRecognizer.
// Since each cv::FaceRecognizer is derived from a
// cv::Algorithm, you can query the data.
//
// First we'll use it to set the threshold of the FaceRecognizer
// to 0.0 without retraining the model. This can be useful if
// you are evaluating the model:
//
model->set("threshold", 0.0);
// Now the threshold of this model is set to 0.0. A prediction
// now returns -1, as it's impossible to have a distance below
// it
predictedLabel = model->predict(testSample);
cout << "Predicted class = " << predictedLabel << endl;
// Show some informations about the model, as there's no cool
// Model data to display as in Eigenfaces/Fisherfaces.
// Due to efficiency reasons the LBP images are not stored
// within the model:
cout << "Model Information:" << endl;
string model_info = format("\tLBPH(radius=%i, neighbors=%i, grid_x=%i, grid_y=%i, threshold=%.2f)",
model->getInt("radius"),
model->getInt("neighbors"),
model->getInt("grid_x"),
model->getInt("grid_y"),
model->getDouble("threshold"));
cout << model_info << endl;
// We could get the histograms for example:
vector<Mat> histograms = model->getMatVector("histograms");
// But should I really visualize it? Probably the length is interesting:
cout << "Size of the histograms: " << histograms[0].total() << endl;
return 0;
}

200
modules/contrib/doc/facerec/src/facerec_save_load.cpp
View File

@@ -1,200 +0,0 @@
/*
* Copyright (c) 2011. Philipp Wagner <bytefish[at]gmx[dot]de>.
* Released to public domain under terms of the BSD Simplified license.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of the organization nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* See <http://www.opensource.org/licenses/bsd-license>
*/
#include "opencv2/contrib.hpp"
#include "opencv2/core.hpp"
#include "opencv2/highgui.hpp"
#include <iostream>
#include <fstream>
#include <sstream>
using namespace cv;
using namespace std;
static Mat norm_0_255(InputArray _src) {
Mat src = _src.getMat();
// Create and return normalized image:
Mat dst;
switch(src.channels()) {
case 1:
cv::normalize(_src, dst, 0, 255, NORM_MINMAX, CV_8UC1);
break;
case 3:
cv::normalize(_src, dst, 0, 255, NORM_MINMAX, CV_8UC3);
break;
default:
src.copyTo(dst);
break;
}
return dst;
}
static void read_csv(const string& filename, vector<Mat>& images, vector<int>& labels, char separator = ';') {
std::ifstream file(filename.c_str(), ifstream::in);
if (!file) {
string error_message = "No valid input file was given, please check the given filename.";
CV_Error(CV_StsBadArg, error_message);
}
string line, path, classlabel;
while (getline(file, line)) {
stringstream liness(line);
getline(liness, path, separator);
getline(liness, classlabel);
if(!path.empty() && !classlabel.empty()) {
images.push_back(imread(path, 0));
labels.push_back(atoi(classlabel.c_str()));
}
}
}
int main(int argc, const char *argv[]) {
// Check for valid command line arguments, print usage
// if no arguments were given.
if (argc < 2) {
cout << "usage: " << argv[0] << " <csv.ext> <output_folder> " << endl;
exit(1);
}
string output_folder = ".";
if (argc == 3) {
output_folder = string(argv[2]);
}
// Get the path to your CSV.
string fn_csv = string(argv[1]);
// These vectors hold the images and corresponding labels.
vector<Mat> images;
vector<int> labels;
// Read in the data. This can fail if no valid
// input filename is given.
try {
read_csv(fn_csv, images, labels);
} catch (cv::Exception& e) {
cerr << "Error opening file \"" << fn_csv << "\". Reason: " << e.msg << endl;
// nothing more we can do
exit(1);
}
// Quit if there are not enough images for this demo.
if(images.size() <= 1) {
string error_message = "This demo needs at least 2 images to work. Please add more images to your data set!";
CV_Error(CV_StsError, error_message);
}
// Get the height from the first image. We'll need this
// later in code to reshape the images to their original
// size:
int height = images[0].rows;
// The following lines simply get the last images from
// your dataset and remove it from the vector. This is
// done, so that the training data (which we learn the
// cv::FaceRecognizer on) and the test data we test
// the model with, do not overlap.
Mat testSample = images[images.size() - 1];
int testLabel = labels[labels.size() - 1];
images.pop_back();
labels.pop_back();
// The following lines create an Eigenfaces model for
// face recognition and train it with the images and
// labels read from the given CSV file.
// This here is a full PCA, if you just want to keep
// 10 principal components (read Eigenfaces), then call
// the factory method like this:
//
// cv::createEigenFaceRecognizer(10);
//
// If you want to create a FaceRecognizer with a
// confidence threshold (e.g. 123.0), call it with:
//
// cv::createEigenFaceRecognizer(10, 123.0);
//
// If you want to use _all_ Eigenfaces and have a threshold,
// then call the method like this:
//
// cv::createEigenFaceRecognizer(0, 123.0);
//
Ptr<FaceRecognizer> model0 = createEigenFaceRecognizer();
model0->train(images, labels);
// save the model to eigenfaces_at.yaml
model0->save("eigenfaces_at.yml");
//
//
// Now create a new Eigenfaces Recognizer
//
Ptr<FaceRecognizer> model1 = createEigenFaceRecognizer();
model1->load("eigenfaces_at.yml");
// The following line predicts the label of a given
// test image:
int predictedLabel = model1->predict(testSample);
//
// To get the confidence of a prediction call the model with:
//
// int predictedLabel = -1;
// double confidence = 0.0;
// model->predict(testSample, predictedLabel, confidence);
//
string result_message = format("Predicted class = %d / Actual class = %d.", predictedLabel, testLabel);
cout << result_message << endl;
// Here is how to get the eigenvalues of this Eigenfaces model:
Mat eigenvalues = model1->getMat("eigenvalues");
// And we can do the same to display the Eigenvectors (read Eigenfaces):
Mat W = model1->getMat("eigenvectors");
// Get the sample mean from the training data
Mat mean = model1->getMat("mean");
// Display or save:
if(argc == 2) {
imshow("mean", norm_0_255(mean.reshape(1, images[0].rows)));
} else {
imwrite(format("%s/mean.png", output_folder.c_str()), norm_0_255(mean.reshape(1, images[0].rows)));
}
// Display or save the Eigenfaces:
for (int i = 0; i < min(10, W.cols); i++) {
string msg = format("Eigenvalue #%d = %.5f", i, eigenvalues.at<double>(i));
cout << msg << endl;
// get eigenvector #i
Mat ev = W.col(i).clone();
// Reshape to original size & normalize to [0...255] for imshow.
Mat grayscale = norm_0_255(ev.reshape(1, height));
// Show the image & apply a Jet colormap for better sensing.
Mat cgrayscale;
applyColorMap(grayscale, cgrayscale, COLORMAP_JET);
// Display or save:
if(argc == 2) {
imshow(format("eigenface_%d", i), cgrayscale);
} else {
imwrite(format("%s/eigenface_%d.png", output_folder.c_str(), i), norm_0_255(cgrayscale));
}
}
// Display or save the image reconstruction at some predefined steps:
for(int num_components = 10; num_components < 300; num_components+=15) {
// slice the eigenvectors from the model
Mat evs = Mat(W, Range::all(), Range(0, num_components));
Mat projection = subspaceProject(evs, mean, images[0].reshape(1,1));
Mat reconstruction = subspaceReconstruct(evs, mean, projection);
// Normalize the result:
reconstruction = norm_0_255(reconstruction.reshape(1, images[0].rows));
// Display or save:
if(argc == 2) {
imshow(format("eigenface_reconstruction_%d", num_components), reconstruction);
} else {
imwrite(format("%s/eigenface_reconstruction_%d.png", output_folder.c_str(), num_components), reconstruction);
}
}
// Display if we are not writing to an output folder:
if(argc == 2) {
waitKey(0);
}
return 0;
}

152
modules/contrib/doc/facerec/src/facerec_video.cpp
View File

@@ -1,152 +0,0 @@
/*
* Copyright (c) 2011. Philipp Wagner <bytefish[at]gmx[dot]de>.
* Released to public domain under terms of the BSD Simplified license.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of the organization nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* See <http://www.opensource.org/licenses/bsd-license>
*/
#include "opencv2/core.hpp"
#include "opencv2/contrib.hpp"
#include "opencv2/highgui.hpp"
#include "opencv2/imgproc.hpp"
#include "opencv2/objdetect.hpp"
#include <iostream>
#include <fstream>
#include <sstream>
using namespace cv;
using namespace std;
static void read_csv(const string& filename, vector<Mat>& images, vector<int>& labels, char separator = ';') {
std::ifstream file(filename.c_str(), ifstream::in);
if (!file) {
string error_message = "No valid input file was given, please check the given filename.";
CV_Error(CV_StsBadArg, error_message);
}
string line, path, classlabel;
while (getline(file, line)) {
stringstream liness(line);
getline(liness, path, separator);
getline(liness, classlabel);
if(!path.empty() && !classlabel.empty()) {
images.push_back(imread(path, 0));
labels.push_back(atoi(classlabel.c_str()));
}
}
}
int main(int argc, const char *argv[]) {
// Check for valid command line arguments, print usage
// if no arguments were given.
if (argc != 4) {
cout << "usage: " << argv[0] << " </path/to/haar_cascade> </path/to/csv.ext> </path/to/device id>" << endl;
cout << "\t </path/to/haar_cascade> -- Path to the Haar Cascade for face detection." << endl;
cout << "\t </path/to/csv.ext> -- Path to the CSV file with the face database." << endl;
cout << "\t <device id> -- The webcam device id to grab frames from." << endl;
exit(1);
}
// Get the path to your CSV:
string fn_haar = string(argv[1]);
string fn_csv = string(argv[2]);
int deviceId = atoi(argv[3]);
// These vectors hold the images and corresponding labels:
vector<Mat> images;
vector<int> labels;
// Read in the data (fails if no valid input filename is given, but you'll get an error message):
try {
read_csv(fn_csv, images, labels);
} catch (cv::Exception& e) {
cerr << "Error opening file \"" << fn_csv << "\". Reason: " << e.msg << endl;
// nothing more we can do
exit(1);
}
// Get the height from the first image. We'll need this
// later in code to reshape the images to their original
// size AND we need to reshape incoming faces to this size:
int im_width = images[0].cols;
int im_height = images[0].rows;
// Create a FaceRecognizer and train it on the given images:
Ptr<FaceRecognizer> model = createFisherFaceRecognizer();
model->train(images, labels);
// That's it for learning the Face Recognition model. You now
// need to create the classifier for the task of Face Detection.
// We are going to use the haar cascade you have specified in the
// command line arguments:
//
CascadeClassifier haar_cascade;
haar_cascade.load(fn_haar);
// Get a handle to the Video device:
VideoCapture cap(deviceId);
// Check if we can use this device at all:
if(!cap.isOpened()) {
cerr << "Capture Device ID " << deviceId << "cannot be opened." << endl;
return -1;
}
// Holds the current frame from the Video device:
Mat frame;
for(;;) {
cap >> frame;
// Clone the current frame:
Mat original = frame.clone();
// Convert the current frame to grayscale:
Mat gray;
cvtColor(original, gray, CV_BGR2GRAY);
// Find the faces in the frame:
vector< Rect_<int> > faces;
haar_cascade.detectMultiScale(gray, faces);
// At this point you have the position of the faces in
// faces. Now we'll get the faces, make a prediction and
// annotate it in the video. Cool or what?
for(int i = 0; i < faces.size(); i++) {
// Process face by face:
Rect face_i = faces[i];
// Crop the face from the image. So simple with OpenCV C++:
Mat face = gray(face_i);
// Resizing the face is necessary for Eigenfaces and Fisherfaces. You can easily
// verify this, by reading through the face recognition tutorial coming with OpenCV.
// Resizing IS NOT NEEDED for Local Binary Patterns Histograms, so preparing the
// input data really depends on the algorithm used.
//
// I strongly encourage you to play around with the algorithms. See which work best
// in your scenario, LBPH should always be a contender for robust face recognition.
//
// Since I am showing the Fisherfaces algorithm here, I also show how to resize the
// face you have just found:
Mat face_resized;
cv::resize(face, face_resized, Size(im_width, im_height), 1.0, 1.0, INTER_CUBIC);
// Now perform the prediction, see how easy that is:
int prediction = model->predict(face_resized);
// And finally write all we've found out to the original image!
// First of all draw a green rectangle around the detected face:
rectangle(original, face_i, CV_RGB(0, 255,0), 1);
// Create the text we will annotate the box with:
string box_text = format("Prediction = %d", prediction);
// Calculate the position for annotated text (make sure we don't
// put illegal values in there):
int pos_x = std::max(face_i.tl().x - 10, 0);
int pos_y = std::max(face_i.tl().y - 10, 0);
// And now put it into the image:
putText(original, box_text, Point(pos_x, pos_y), FONT_HERSHEY_PLAIN, 1.0, CV_RGB(0,255,0), 2.0);
}
// Show the result:
imshow("face_recognizer", original);
// And display it:
char key = (char) waitKey(20);
// Exit this loop on escape:
if(key == 27)
break;
}
return 0;
}

233
modules/contrib/doc/facerec/tutorial/facerec_gender_classification.rst
View File

@@ -1,233 +0,0 @@
Gender Classification with OpenCV
=================================
.. contents:: Table of Contents
:depth: 3
Introduction
------------
A lot of people interested in face recognition, also want to know how to perform image classification tasks like:
* Gender Classification (Gender Detection)
* Emotion Classification (Emotion Detection)
* Glasses Classification (Glasses Detection)
* ...
This is has become very, very easy with the new :ocv:class:`FaceRecognizer` class. In this tutorial I'll show you how to perform gender classification with OpenCV on a set of face images. You'll also learn how to align your images to enhance the recognition results. If you want to do emotion classification instead of gender classification, all you need to do is to update is your training data and the configuration you pass to the demo.
Prerequisites
--------------
For gender classification of faces, you'll need some images of male and female faces first. I've decided to search faces of celebrities using `Google Images <http://www.google.com/images>`_ with the faces filter turned on (my god, they have great algorithms at `Google <http://www.google.com>`_!). My database has 8 male and 5 female subjects, each with 10 images. Here are the names, if you don't know who to search:
* Angelina Jolie
* Arnold Schwarzenegger
* Brad Pitt
* Emma Watson
* George Clooney
* Jennifer Lopez
* Johnny Depp
* Justin Timberlake
* Katy Perry
* Keanu Reeves
* Naomi Watts
* Patrick Stewart
* Tom Cruise
Once you have acquired some images, you'll need to read them. In the demo application I have decided to read the images from a very simple CSV file. Why? Because it's the simplest platform-independent approach I can think of. However, if you know a simpler solution please ping me about it. Basically all the CSV file needs to contain are lines composed of a ``filename`` followed by a ``;`` followed by the ``label`` (as *integer number*), making up a line like this:
.. code-block:: none
/path/to/image.ext;0
Let's dissect the line. ``/path/to/image.ext`` is the path to an image, probably something like this if you are in Windows: ``C:/faces/person0/image0.jpg``. Then there is the separator ``;`` and finally we assign a label ``0`` to the image. Think of the label as the subject (the person, the gender or whatever comes to your mind). In the gender classification scenario, the label is the gender the person has. I'll give the label ``0`` to *male* persons and the label ``1`` is for *female* subjects. So my CSV file looks like this:
.. code-block:: none
/home/philipp/facerec/data/gender/male/keanu_reeves/keanu_reeves_01.jpg;0
/home/philipp/facerec/data/gender/male/keanu_reeves/keanu_reeves_02.jpg;0
/home/philipp/facerec/data/gender/male/keanu_reeves/keanu_reeves_03.jpg;0
...
/home/philipp/facerec/data/gender/female/katy_perry/katy_perry_01.jpg;1
/home/philipp/facerec/data/gender/female/katy_perry/katy_perry_02.jpg;1
/home/philipp/facerec/data/gender/female/katy_perry/katy_perry_03.jpg;1
...
/home/philipp/facerec/data/gender/male/brad_pitt/brad_pitt_01.jpg;0
/home/philipp/facerec/data/gender/male/brad_pitt/brad_pitt_02.jpg;0
/home/philipp/facerec/data/gender/male/brad_pitt/brad_pitt_03.jpg;0
...
/home/philipp/facerec/data/gender/female/emma_watson/emma_watson_08.jpg;1
/home/philipp/facerec/data/gender/female/emma_watson/emma_watson_02.jpg;1
/home/philipp/facerec/data/gender/female/emma_watson/emma_watson_03.jpg;1
All images for this example were chosen to have a frontal face perspective. They have been cropped, scaled and rotated to be aligned at the eyes, just like this set of George Clooney images:
.. image:: ../img/tutorial/gender_classification/clooney_set.png
:align: center
You really don't want to create the CSV file by hand. And you really don't want scale, rotate & translate the images manually. I have prepared you two Python scripts ``create_csv.py`` and ``crop_face.py``, you can find them in the ``src`` folder coming with this documentation. You'll see how to use them in the :ref:`appendixfgc`.
Fisherfaces for Gender Classification
--------------------------------------
If you want to decide whether a person is *male* or *female*, you have to learn the discriminative features of both classes. The Eigenfaces method is based on the Principal Component Analysis, which is an unsupervised statistical model and not suitable for this task. Please see the Face Recognition tutorial for insights into the algorithms. The Fisherfaces instead yields a class-specific linear projection, so it is much better suited for the gender classification task. `http://www.bytefish.de/blog/gender_classification <http://www.bytefish.de/blog/gender_classification>`_ shows the recognition rate of the Fisherfaces method for gender classification.
The Fisherfaces method achieves a 98% recognition rate in a subject-independent cross-validation. A subject-independent cross-validation means *images of the person under test are never used for learning the model*. And could you believe it: you can simply use the facerec_fisherfaces demo, that's inlcuded in OpenCV.
Fisherfaces in OpenCV
---------------------
The source code for this demo application is also available in the ``src`` folder coming with this documentation:
* :download:`src/facerec_fisherfaces.cpp <../src/facerec_fisherfaces.cpp>`
.. literalinclude:: ../src/facerec_fisherfaces.cpp
:language: cpp
:linenos:
Running the Demo
----------------
If you are in Windows, then simply start the demo by running (from command line):
.. code-block:: none
facerec_fisherfaces.exe C:/path/to/your/csv.ext
If you are in Linux, then simply start the demo by running:
.. code-block:: none
./facerec_fisherfaces /path/to/your/csv.ext
If you don't want to display the images, but save them, then pass the desired path to the demo. It works like this in Windows:
.. code-block:: none
facerec_fisherfaces.exe C:/path/to/your/csv.ext C:/path/to/store/results/at
And in Linux:
.. code-block:: none
./facerec_fisherfaces /path/to/your/csv.ext /path/to/store/results/at
Results
-------
If you run the program with your CSV file as parameter, you'll see the Fisherface that separates between male and female images. I've decided to apply a Jet colormap in this demo, so you can see which features the method identifies:
.. image:: ../img/tutorial/gender_classification/fisherface_0.png
The demo also shows the average face of the male and female training images you have passed:
.. image:: ../img/tutorial/gender_classification/mean.png
Moreover it the demo should yield the prediction for the correct gender:
.. code-block:: none
Predicted class = 1 / Actual class = 1.
And for advanced users I have also shown the Eigenvalue for the Fisherface:
.. code-block:: none
Eigenvalue #0 = 152.49493
And the Fisherfaces reconstruction:
.. image:: ../img/tutorial/gender_classification/fisherface_reconstruction_0.png
I hope this gives you an idea how to approach gender classification and the other image classification tasks.
.. _appendixfgc:
Appendix
--------
Creating the CSV File
+++++++++++++++++++++
You don't really want to create the CSV file by hand. I have prepared you a little Python script ``create_csv.py`` (you find it at ``/src/create_csv.py`` coming with this tutorial) that automatically creates you a CSV file. If you have your images in hierarchie like this (``/basepath/<subject>/<image.ext>``):
.. code-block:: none
philipp@mango:~/facerec/data/at$ tree
.
|-- s1
| |-- 1.pgm
| |-- ...
| |-- 10.pgm
|-- s2
| |-- 1.pgm
| |-- ...
| |-- 10.pgm
...
|-- s40
| |-- 1.pgm
| |-- ...
| |-- 10.pgm
Then simply call ``create_csv.py`` with the path to the folder, just like this and you could save the output:
.. code-block:: none
philipp@mango:~/facerec/data$ python create_csv.py
at/s13/2.pgm;0
at/s13/7.pgm;0
at/s13/6.pgm;0
at/s13/9.pgm;0
at/s13/5.pgm;0
at/s13/3.pgm;0
at/s13/4.pgm;0
at/s13/10.pgm;0
at/s13/8.pgm;0
at/s13/1.pgm;0
at/s17/2.pgm;1
at/s17/7.pgm;1
at/s17/6.pgm;1
at/s17/9.pgm;1
at/s17/5.pgm;1
at/s17/3.pgm;1
[...]
Here is the script, if you can't find it:
.. literalinclude:: ../src/create_csv.py
:language: python
:linenos:
Aligning Face Images
++++++++++++++++++++
An accurate alignment of your image data is especially important in tasks like emotion detection, were you need as much detail as possible. Believe me... You don't want to do this by hand. So I've prepared you a tiny Python script. The code is really easy to use. To scale, rotate and crop the face image you just need to call *CropFace(image, eye_left, eye_right, offset_pct, dest_sz)*, where:
* *eye_left* is the position of the left eye
* *eye_right* is the position of the right eye
* *offset_pct* is the percent of the image you want to keep next to the eyes (horizontal, vertical direction)
* *dest_sz* is the size of the output image
If you are using the same *offset_pct* and *dest_sz* for your images, they are all aligned at the eyes.
.. literalinclude:: ../src/crop_face.py
:language: python
:linenos:
Imagine we are given `this photo of Arnold Schwarzenegger <http://en.wikipedia.org/wiki/File:Arnold_Schwarzenegger_edit%28ws%29.jpg>`_, which is under a Public Domain license. The (x,y)-position of the eyes is approximately *(252,364)* for the left and *(420,366)* for the right eye. Now you only need to define the horizontal offset, vertical offset and the size your scaled, rotated & cropped face should have.
Here are some examples:
+---------------------------------+----------------------------------------------------------------------------+
| Configuration | Cropped, Scaled, Rotated Face |
+=================================+============================================================================+
| 0.1 (10%), 0.1 (10%), (200,200) | .. image:: ../img/tutorial/gender_classification/arnie_10_10_200_200.jpg |
+---------------------------------+----------------------------------------------------------------------------+
| 0.2 (20%), 0.2 (20%), (200,200) | .. image:: ../img/tutorial/gender_classification/arnie_20_20_200_200.jpg |
+---------------------------------+----------------------------------------------------------------------------+
| 0.3 (30%), 0.3 (30%), (200,200) | .. image:: ../img/tutorial/gender_classification/arnie_30_30_200_200.jpg |
+---------------------------------+----------------------------------------------------------------------------+
| 0.2 (20%), 0.2 (20%), (70,70) | .. image:: ../img/tutorial/gender_classification/arnie_20_20_70_70.jpg |
+---------------------------------+----------------------------------------------------------------------------+

46
modules/contrib/doc/facerec/tutorial/facerec_save_load.rst
View File

@@ -1,46 +0,0 @@
Saving and Loading a FaceRecognizer
===================================
Introduction
------------
Saving and loading a :ocv:class:`FaceRecognizer` is very important. Training a FaceRecognizer can be a very time-intense task, plus it's often impossible to ship the whole face database to the user of your product. The task of saving and loading a FaceRecognizer is easy with :ocv:class:`FaceRecognizer`. You only have to call :ocv:func:`FaceRecognizer::load` for loading and :ocv:func:`FaceRecognizer::save` for saving a :ocv:class:`FaceRecognizer`.
I'll adapt the Eigenfaces example from the :doc:`../facerec_tutorial`: Imagine we want to learn the Eigenfaces of the `AT&T Facedatabase <http://www.cl.cam.ac.uk/research/dtg/attarchive/facedatabase.html>`_, store the model to a YAML file and then load it again.
From the loaded model, we'll get a prediction, show the mean, Eigenfaces and the image reconstruction.
Using FaceRecognizer::save and FaceRecognizer::load
-----------------------------------------------------
The source code for this demo application is also available in the ``src`` folder coming with this documentation:
* :download:`src/facerec_save_load.cpp <../src/facerec_save_load.cpp>`
.. literalinclude:: ../src/facerec_save_load.cpp
:language: cpp
:linenos:
Results
-------
``eigenfaces_at.yml`` then contains the model state, we'll simply look at the first 10 lines with ``head eigenfaces_at.yml``:
.. code-block:: none
philipp@mango:~/github/libfacerec-build$ head eigenfaces_at.yml
%YAML:1.0
num_components: 399
mean: !!opencv-matrix
rows: 1
cols: 10304
dt: d
data: [ 8.5558897243107765e+01, 8.5511278195488714e+01,
8.5854636591478695e+01, 8.5796992481203006e+01,
8.5952380952380949e+01, 8.6162907268170414e+01,
8.6082706766917283e+01, 8.5776942355889716e+01,
And here is the Reconstruction, which is the same as the original:
.. image:: ../img/eigenface_reconstruction_opencv.png
:align: center

207
modules/contrib/doc/facerec/tutorial/facerec_video_recognition.rst
View File

@@ -1,207 +0,0 @@
Face Recognition in Videos with OpenCV
=======================================
.. contents:: Table of Contents
:depth: 3
Introduction
------------
Whenever you hear the term *face recognition*, you instantly think of surveillance in videos. So performing face recognition in videos (e.g. webcam) is one of the most requested features I have got. I have heard your cries, so here it is. An application, that shows you how to do face recognition in videos! For the face detection part we'll use the awesome :ocv:class:`CascadeClassifier` and we'll use :ocv:class:`FaceRecognizer` for face recognition. This example uses the Fisherfaces method for face recognition, because it is robust against large changes in illumination.
Here is what the final application looks like. As you can see I am only writing the id of the recognized person above the detected face (by the way this id is Arnold Schwarzenegger for my data set):
.. image:: ../img/tutorial/facerec_video/facerec_video.png
:align: center
:scale: 70%
This demo is a basis for your research and it shows you how to implement face recognition in videos. You probably want to extend the application and make it more sophisticated: You could combine the id with the name, then show the confidence of the prediction, recognize the emotion... and and and. But before you send mails, asking what these Haar-Cascade thing is or what a CSV is: Make sure you have read the entire tutorial. It's all explained in here. If you just want to scroll down to the code, please note:
* The available Haar-Cascades for face detection are located in the ``data`` folder of your OpenCV installation! One of the available Haar-Cascades for face detection is for example ``/path/to/opencv/data/haarcascades/haarcascade_frontalface_default.xml``.
I encourage you to experiment with the application. Play around with the available :ocv:class:`FaceRecognizer` implementations, try the available cascades in OpenCV and see if you can improve your results!
Prerequisites
--------------
You want to do face recognition, so you need some face images to learn a :ocv:class:`FaceRecognizer` on. I have decided to reuse the images from the gender classification example: :doc:`facerec_gender_classification`.
I have the following celebrities in my training data set:
* Angelina Jolie
* Arnold Schwarzenegger
* Brad Pitt
* George Clooney
* Johnny Depp
* Justin Timberlake
* Katy Perry
* Keanu Reeves
* Patrick Stewart
* Tom Cruise
In the demo I have decided to read the images from a very simple CSV file. Why? Because it's the simplest platform-independent approach I can think of. However, if you know a simpler solution please ping me about it. Basically all the CSV file needs to contain are lines composed of a ``filename`` followed by a ``;`` followed by the ``label`` (as *integer number*), making up a line like this:
.. code-block:: none
/path/to/image.ext;0
Let's dissect the line. ``/path/to/image.ext`` is the path to an image, probably something like this if you are in Windows: ``C:/faces/person0/image0.jpg``. Then there is the separator ``;`` and finally we assign a label ``0`` to the image. Think of the label as the subject (the person, the gender or whatever comes to your mind). In the face recognition scenario, the label is the person this image belongs to. In the gender classification scenario, the label is the gender the person has. So my CSV file looks like this:
.. code-block:: none
/home/philipp/facerec/data/c/keanu_reeves/keanu_reeves_01.jpg;0
/home/philipp/facerec/data/c/keanu_reeves/keanu_reeves_02.jpg;0
/home/philipp/facerec/data/c/keanu_reeves/keanu_reeves_03.jpg;0
...
/home/philipp/facerec/data/c/katy_perry/katy_perry_01.jpg;1
/home/philipp/facerec/data/c/katy_perry/katy_perry_02.jpg;1
/home/philipp/facerec/data/c/katy_perry/katy_perry_03.jpg;1
...
/home/philipp/facerec/data/c/brad_pitt/brad_pitt_01.jpg;2
/home/philipp/facerec/data/c/brad_pitt/brad_pitt_02.jpg;2
/home/philipp/facerec/data/c/brad_pitt/brad_pitt_03.jpg;2
...
/home/philipp/facerec/data/c1/crop_arnold_schwarzenegger/crop_08.jpg;6
/home/philipp/facerec/data/c1/crop_arnold_schwarzenegger/crop_05.jpg;6
/home/philipp/facerec/data/c1/crop_arnold_schwarzenegger/crop_02.jpg;6
/home/philipp/facerec/data/c1/crop_arnold_schwarzenegger/crop_03.jpg;6
All images for this example were chosen to have a frontal face perspective. They have been cropped, scaled and rotated to be aligned at the eyes, just like this set of George Clooney images:
.. image:: ../img/tutorial/gender_classification/clooney_set.png
:align: center
Face Recongition from Videos
-----------------------------
The source code for the demo is available in the ``src`` folder coming with this documentation:
* :download:`src/facerec_video.cpp <../src/facerec_video.cpp>`
This demo uses the :ocv:class:`CascadeClassifier`:
.. literalinclude:: ../src/facerec_video.cpp
:language: cpp
:linenos:
Running the Demo
----------------
You'll need:
* The path to a valid Haar-Cascade for detecting a face with a :ocv:class:`CascadeClassifier`.
* The path to a valid CSV File for learning a :ocv:class:`FaceRecognizer`.
* A webcam and its device id (you don't know the device id? Simply start from 0 on and see what happens).
If you are in Windows, then simply start the demo by running (from command line):
.. code-block:: none
facerec_video.exe <C:/path/to/your/haar_cascade.xml> <C:/path/to/your/csv.ext> <video device>
If you are in Linux, then simply start the demo by running:
.. code-block:: none
./facerec_video </path/to/your/haar_cascade.xml> </path/to/your/csv.ext> <video device>
An example. If the haar-cascade is at ``C:/opencv/data/haarcascades/haarcascade_frontalface_default.xml``, the CSV file is at ``C:/facerec/data/celebrities.txt`` and I have a webcam with deviceId ``1``, then I would call the demo with:
.. code-block:: none
facerec_video.exe C:/opencv/data/haarcascades/haarcascade_frontalface_default.xml C:/facerec/data/celebrities.txt 1
That's it.
Results
-------
Enjoy!
Appendix
--------
Creating the CSV File
+++++++++++++++++++++
You don't really want to create the CSV file by hand. I have prepared you a little Python script ``create_csv.py`` (you find it at ``/src/create_csv.py`` coming with this tutorial) that automatically creates you a CSV file. If you have your images in hierarchie like this (``/basepath/<subject>/<image.ext>``):
.. code-block:: none
philipp@mango:~/facerec/data/at$ tree
.
|-- s1
| |-- 1.pgm
| |-- ...
| |-- 10.pgm
|-- s2
| |-- 1.pgm
| |-- ...
| |-- 10.pgm
...
|-- s40
| |-- 1.pgm
| |-- ...
| |-- 10.pgm
Then simply call ``create_csv.py`` with the path to the folder, just like this and you could save the output:
.. code-block:: none
philipp@mango:~/facerec/data$ python create_csv.py
at/s13/2.pgm;0
at/s13/7.pgm;0
at/s13/6.pgm;0
at/s13/9.pgm;0
at/s13/5.pgm;0
at/s13/3.pgm;0
at/s13/4.pgm;0
at/s13/10.pgm;0
at/s13/8.pgm;0
at/s13/1.pgm;0
at/s17/2.pgm;1
at/s17/7.pgm;1
at/s17/6.pgm;1
at/s17/9.pgm;1
at/s17/5.pgm;1
at/s17/3.pgm;1
[...]
Here is the script, if you can't find it:
.. literalinclude:: ../src/create_csv.py
:language: python
:linenos:
Aligning Face Images
++++++++++++++++++++
An accurate alignment of your image data is especially important in tasks like emotion detection, were you need as much detail as possible. Believe me... You don't want to do this by hand. So I've prepared you a tiny Python script. The code is really easy to use. To scale, rotate and crop the face image you just need to call *CropFace(image, eye_left, eye_right, offset_pct, dest_sz)*, where:
* *eye_left* is the position of the left eye
* *eye_right* is the position of the right eye
* *offset_pct* is the percent of the image you want to keep next to the eyes (horizontal, vertical direction)
* *dest_sz* is the size of the output image
If you are using the same *offset_pct* and *dest_sz* for your images, they are all aligned at the eyes.
.. literalinclude:: ../src/crop_face.py
:language: python
:linenos:
Imagine we are given `this photo of Arnold Schwarzenegger <http://en.wikipedia.org/wiki/File:Arnold_Schwarzenegger_edit%28ws%29.jpg>`_, which is under a Public Domain license. The (x,y)-position of the eyes is approximately *(252,364)* for the left and *(420,366)* for the right eye. Now you only need to define the horizontal offset, vertical offset and the size your scaled, rotated & cropped face should have.
Here are some examples:
+---------------------------------+----------------------------------------------------------------------------+
| Configuration | Cropped, Scaled, Rotated Face |
+=================================+============================================================================+
| 0.1 (10%), 0.1 (10%), (200,200) | .. image:: ../img/tutorial/gender_classification/arnie_10_10_200_200.jpg |
+---------------------------------+----------------------------------------------------------------------------+
| 0.2 (20%), 0.2 (20%), (200,200) | .. image:: ../img/tutorial/gender_classification/arnie_20_20_200_200.jpg |
+---------------------------------+----------------------------------------------------------------------------+
| 0.3 (30%), 0.3 (30%), (200,200) | .. image:: ../img/tutorial/gender_classification/arnie_30_30_200_200.jpg |
+---------------------------------+----------------------------------------------------------------------------+
| 0.2 (20%), 0.2 (20%), (70,70) | .. image:: ../img/tutorial/gender_classification/arnie_20_20_70_70.jpg |
+---------------------------------+----------------------------------------------------------------------------+

232
modules/contrib/doc/openfabmap.rst
View File

@@ -1,232 +0,0 @@
OpenFABMAP
========================================
.. highlight:: cpp
The openFABMAP package has been integrated into OpenCV from the openFABMAP <http://code.google.com/p/openfabmap/> project [ICRA2011]_. OpenFABMAP is an open and modifiable code-source which implements the Fast Appearance-based Mapping algorithm (FAB-MAP) developed by Mark Cummins and Paul Newman. The algorithms used in openFABMAP were developed using only the relevant FAB-MAP publications.
FAB-MAP is an approach to appearance-based place recognition. FAB-MAP compares images of locations that have been visited and determines the probability of re-visiting a location, as well as providing a measure of the probability of being at a new, previously unvisited location. Camera images form the sole input to the system, from which visual bag-of-words models are formed through the extraction of appearance-based (e.g. SURF) features.
openFABMAP requires training data (e.g. a collection of images from a similar but not identical environment) to construct a visual vocabulary for the visual bag-of-words model, along with a Chow-Liu tree representation of feature likelihood and for use in the Sampled new place method (see below).
.. note::
* An example using the openFABMAP package can be found at opencv_source_code/samples/cpp/fabmap_sample.cpp
of2::FabMap
--------------------
.. ocv:class:: of2::FabMap
The main FabMap class performs the comparison between visual bags-of-words extracted from one or more images. The FabMap class is instantiated as one of the four inherited FabMap classes (FabMap1, FabMapLUT, FabMapFBO, FabMap2). Each inherited class performs the comparison differently based on algorithm iterations as published (see each class below for specifics). A Chow-Liu tree, detector model parameters and some option flags are common to all Fabmap variants and are supplied on class creation. Training data (visual bag-of-words) is supplied to the class if using the SAMPLED new place method. Test data (visual bag-of-words) is supplied as images to which query bag-of-words are compared against. The common flags are listed below: ::
enum {
MEAN_FIELD,
SAMPLED,
NAIVE_BAYES,
CHOW_LIU,
MOTION_MODEL
};
#. MEAN_FIELD: Use the Mean Field approximation to determine the new place likelihood (cannot be used for FabMap2).
#. SAMPLED: Use the Sampled approximation to determine the new place likelihood. Requires training data (see below).
#. NAIVE_BAYES: Assume a naive Bayes approximation to feature distribution (i.e. all features are independent). Note that a Chow-Liu tree is still required but only the absolute word probabilities are used, feature co-occurrance information is discarded.
#. CHOW_LIU: Use the full Chow-Liu tree to approximate feature distribution.
#. MOTION_MODEL: Update the location distribution using the previous distribution as a (weak) prior. Used for matching in sequences (i.e. successive video frames).
Training Data
++++++++++++++++++++
Training data is required to use the SAMPLED new place method. The SAMPLED method was shown to have improved performance over the alternative MEAN_FIELD method. Training data can be added singularly or as a batch.
.. ocv:function:: virtual void addTraining(const Mat& queryImgDescriptor)
:param queryImgDescriptor: bag-of-words image descriptors stored as rows in a Mat
.. ocv:function:: virtual void addTraining(const vector<Mat>& queryImgDescriptors)
:param queryImgDescriptors: a vector containing multiple bag-of-words image descriptors
.. ocv:function:: const vector<Mat>& getTrainingImgDescriptors() const
Returns a vector containing multiple bag-of-words image descriptors
Test Data
++++++++++++++++++++
Test Data is the database of images represented using bag-of-words models. When a compare function is called, each query point is compared to the test data.
.. ocv:function:: virtual void add(const Mat& queryImgDescriptor)
:param queryImgDescriptor: bag-of-words image descriptors stored as rows in a Mat
.. ocv:function:: virtual void add(const vector<Mat>& queryImgDescriptors)
:param queryImgDescriptors: a vector containing multiple bag-of-words image descriptors
.. ocv:function:: const vector<Mat>& getTestImgDescriptors() const
Returns a vector containing multiple bag-of-words image descriptors
Image Comparison
++++++++++++++++++++
Image matching is performed calling the compare function. Query bag-of-words image descriptors are provided and compared to test data added to the FabMap class. Alternatively test data can be provided with the call to compare to which the comparison is performed. Results are written to the 'matches' argument.
.. ocv:function:: void compare(const Mat& queryImgDescriptor, vector<IMatch>& matches, bool addQuery = false, const Mat& mask = Mat())
:param queryImgDescriptor: bag-of-words image descriptors stored as rows in a Mat
:param matches: a vector of image match probabilities
:param addQuery: if true the queryImg Descriptor is added to the test data after the comparison is performed.
:param mask: *not implemented*
.. ocv:function:: void compare(const Mat& queryImgDescriptor, const Mat& testImgDescriptors, vector<IMatch>& matches, const Mat& mask = Mat())
:param testImgDescriptors: bag-of-words image descriptors stored as rows in a Mat
.. ocv:function:: void compare(const Mat& queryImgDescriptor, const vector<Mat>& testImgDescriptors, vector<IMatch>& matches, const Mat& mask = Mat())
:param testImgDescriptors: a vector of multiple bag-of-words image descriptors
.. ocv:function:: void compare(const vector<Mat>& queryImgDescriptors, vector<IMatch>& matches, bool addQuery = false, const Mat& mask = Mat())
:param queryImgDescriptors: a vector of multiple bag-of-words image descriptors
.. ocv:function:: void compare(const vector<Mat>& queryImgDescriptors, const vector<Mat>& testImgDescriptors, vector<IMatch>& matches, const Mat& mask = Mat())
FabMap classes
++++++++++++++++++++
.. ocv:class:: FabMap1 : public FabMap
The original FAB-MAP algorithm without any computational improvements as published in [IJRR2008]_
.. ocv:function:: FabMap1::FabMap1(const Mat& clTree, double PzGe, double PzGNe, int flags, int numSamples = 0)
:param clTree: a Chow-Liu tree class
:param PzGe: the dector model recall. The probability of the feature detector extracting a feature from an object given it is in the scene. This is used to account for detector noise.
:param PzGNe: the dector model precision. The probability of the feature detector falsing extracting a feature representing an object that is not in the scene.
:param numSamples: the number of samples to use for the SAMPLED new place calculation
.. ocv:class:: FabMapLUT : public FabMap
The original FAB-MAP algorithm implemented as a look-up table for speed enhancements [ICRA2011]_
.. ocv:function:: FabMapLUT::FabMapLUT(const Mat& clTree, double PzGe, double PzGNe, int flags, int numSamples = 0, int precision = 6)
:param precision: the precision with which to store the pre-computed likelihoods
.. ocv:class:: FabMapFBO : public FabMap
The accelerated FAB-MAP using a 'fast bail-out' approach as in [TRO2010]_
.. ocv:function:: FabMapFBO::FabMapFBO(const Mat& clTree, double PzGe, double PzGNe, int flags, int numSamples = 0, double rejectionThreshold = 1e-8, double PsGd = 1e-8, int bisectionStart = 512, int bisectionIts = 9)
:param rejectionThreshold: images are not considered a match when the likelihood falls below the Bennett bound by the amount given by the rejectionThreshold. The threshold provides a speed/accuracy trade-off. A lower bound will be more accurate
:param PsGd: used to calculate the Bennett bound. Provides a speed/accuracy trade-off. A lower bound will be more accurate
:param bisectionStart: Used to estimate the bound using the bisection method. Must be larger than the largest expected difference between maximum and minimum image likelihoods
:param bisectionIts: The number of iterations for which to perform the bisection method
.. ocv:class:: FabMap2 : public FabMap
The inverted index FAB-MAP as in [IJRR2010]_. This version of FAB-MAP is the fastest without any loss of accuracy.
.. ocv:function:: FabMap2::FabMap2(const Mat& clTree, double PzGe, double PzGNe, int flags)
.. [IJRR2008] M. Cummins and P. Newman, "FAB-MAP: Probabilistic Localization and Mapping in the Space of Appearance," The International Journal of Robotics Research, vol. 27(6), pp. 647-665, 2008
.. [TRO2010] M. Cummins and P. Newman, "Accelerating FAB-MAP with concentration inequalities," IEEE Transactions on Robotics, vol. 26(6), pp. 1042-1050, 2010
.. [IJRR2010] M. Cummins and P. Newman, "Appearance-only SLAM at large scale with FAB-MAP 2.0," The International Journal of Robotics Research, vol. 30(9), pp. 1100-1123, 2010
.. [ICRA2011] A. Glover, et al., "OpenFABMAP: An Open Source Toolbox for Appearance-based Loop Closure Detection," in IEEE International Conference on Robotics and Automation, St Paul, Minnesota, 2011
of2::IMatch
--------------------
.. ocv:struct:: of2::IMatch
FAB-MAP comparison results are stored in a vector of IMatch structs. Each IMatch structure provides the index of the provided query bag-of-words, the index of the test bag-of-words, the raw log-likelihood of the match (independent of other comparisons), and the match probability (normalised over other comparison likelihoods).
::
struct IMatch {
IMatch() :
queryIdx(-1), imgIdx(-1), likelihood(-DBL_MAX), match(-DBL_MAX) {
}
IMatch(int _queryIdx, int _imgIdx, double _likelihood, double _match) :
queryIdx(_queryIdx), imgIdx(_imgIdx), likelihood(_likelihood), match(
_match) {
}
int queryIdx; //query index
int imgIdx; //test index
double likelihood; //raw loglikelihood
double match; //normalised probability
bool operator<(const IMatch& m) const {
return match < m.match;
}
};
of2::ChowLiuTree
--------------------
.. ocv:class:: of2::ChowLiuTree
The Chow-Liu tree is a probabilistic model of the environment in terms of feature occurance and co-occurance. The Chow-Liu tree is a form of Bayesian network. FAB-MAP uses the model when calculating bag-of-words similarity by taking into account feature saliency. Training data is provided to the ChowLiuTree class in the form of bag-of-words image descriptors. The make function produces a cv::Mat that encodes the tree structure.
.. ocv:function:: of2::ChowLiuTree::ChowLiuTree()
.. ocv:function:: void of2::ChowLiuTree::add(const Mat& imgDescriptor)
:param imgDescriptor: bag-of-words image descriptors stored as rows in a Mat
.. ocv:function:: void of2::ChowLiuTree::add(const vector<Mat>& imgDescriptors)
:param imgDescriptors: a vector containing multiple bag-of-words image descriptors
.. ocv:function:: const vector<Mat>& of2::ChowLiuTree::getImgDescriptors() const
Returns a vector containing multiple bag-of-words image descriptors
.. ocv:function:: Mat of2::ChowLiuTree::make(double infoThreshold = 0.0)
:param infoThreshold: a threshold can be set to reduce the amount of memory used when making the Chow-Liu tree, which can occur with large vocabulary sizes. This function can fail if the threshold is set too high. If memory is an issue the value must be set by trial and error (~0.0005)
of2::BOWMSCTrainer
--------------------
.. ocv:class:: of2::BOWMSCTrainer : public of2::BOWTrainer
BOWMSCTrainer is a custom clustering algorithm used to produce the feature vocabulary required to create bag-of-words representations. The algorithm is an implementation of [AVC2007]_. Arguments against using K-means for the FAB-MAP algorithm are discussed in [IJRR2010]_. The BOWMSCTrainer inherits from the cv::BOWTrainer class, overwriting the cluster function.
.. ocv:function:: of2::BOWMSCTrainer::BOWMSCTrainer(double clusterSize = 0.4)
:param clusterSize: the specificity of the vocabulary produced. A smaller cluster size will instigate a larger vocabulary.
.. ocv:function:: virtual Mat of2::BOWMSCTrainer::cluster() const
Cluster using features added to the class
.. ocv:function:: virtual Mat of2::BOWMSCTrainer::cluster(const Mat& descriptors) const
:param descriptors: feature descriptors provided as rows of the Mat.
.. [AVC2007] Alexandra Teynor and Hans Burkhardt, "Fast Codebook Generation by Sequential Data Analysis for Object Classification", in Advances in Visual Computing, pp. 610-620, 2007

115
modules/contrib/doc/stereo.rst
View File

@@ -1,115 +0,0 @@
Stereo Correspondence
========================================
.. highlight:: cpp
StereoVar
----------
.. ocv:class:: StereoVar
Class for computing stereo correspondence using the variational matching algorithm ::
class StereoVar
{
StereoVar();
StereoVar( int levels, double pyrScale,
int nIt, int minDisp, int maxDisp,
int poly_n, double poly_sigma, float fi,
float lambda, int penalization, int cycle,
int flags);
virtual ~StereoVar();
virtual void operator()(InputArray left, InputArray right, OutputArray disp);
int levels;
double pyrScale;
int nIt;
int minDisp;
int maxDisp;
int poly_n;
double poly_sigma;
float fi;
float lambda;
int penalization;
int cycle;
int flags;
...
};
The class implements the modified S. G. Kosov algorithm [Publication] that differs from the original one as follows:
* The automatic initialization of method's parameters is added.
* The method of Smart Iteration Distribution (SID) is implemented.
* The support of Multi-Level Adaptation Technique (MLAT) is not included.
* The method of dynamic adaptation of method's parameters is not included.
StereoVar::StereoVar
--------------------------
.. ocv:function:: StereoVar::StereoVar()
.. ocv:function:: StereoVar::StereoVar( int levels, double pyrScale, int nIt, int minDisp, int maxDisp, int poly_n, double poly_sigma, float fi, float lambda, int penalization, int cycle, int flags )
The constructor
:param levels: The number of pyramid layers, including the initial image. levels=1 means that no extra layers are created and only the original images are used. This parameter is ignored if flag USE_AUTO_PARAMS is set.
:param pyrScale: Specifies the image scale (<1) to build the pyramids for each image. pyrScale=0.5 means the classical pyramid, where each next layer is twice smaller than the previous. (This parameter is ignored if flag USE_AUTO_PARAMS is set).
:param nIt: The number of iterations the algorithm does at each pyramid level. (If the flag USE_SMART_ID is set, the number of iterations will be redistributed in such a way, that more iterations will be done on more coarser levels.)
:param minDisp: Minimum possible disparity value. Could be negative in case the left and right input images change places.
:param maxDisp: Maximum possible disparity value.
:param poly_n: Size of the pixel neighbourhood used to find polynomial expansion in each pixel. The larger values mean that the image will be approximated with smoother surfaces, yielding more robust algorithm and more blurred motion field. Typically, poly_n = 3, 5 or 7
:param poly_sigma: Standard deviation of the Gaussian that is used to smooth derivatives that are used as a basis for the polynomial expansion. For poly_n=5 you can set poly_sigma=1.1 , for poly_n=7 a good value would be poly_sigma=1.5
:param fi: The smoothness parameter, ot the weight coefficient for the smoothness term.
:param lambda: The threshold parameter for edge-preserving smoothness. (This parameter is ignored if PENALIZATION_CHARBONNIER or PENALIZATION_PERONA_MALIK is used.)
:param penalization: Possible values: PENALIZATION_TICHONOV - linear smoothness; PENALIZATION_CHARBONNIER - non-linear edge preserving smoothness; PENALIZATION_PERONA_MALIK - non-linear edge-enhancing smoothness. (This parameter is ignored if flag USE_AUTO_PARAMS is set).
:param cycle: Type of the multigrid cycle. Possible values: CYCLE_O and CYCLE_V for null- and v-cycles respectively. (This parameter is ignored if flag USE_AUTO_PARAMS is set).
:param flags: The operation flags; can be a combination of the following:
* USE_INITIAL_DISPARITY: Use the input flow as the initial flow approximation.
* USE_EQUALIZE_HIST: Use the histogram equalization in the pre-processing phase.
* USE_SMART_ID: Use the smart iteration distribution (SID).
* USE_AUTO_PARAMS: Allow the method to initialize the main parameters.
* USE_MEDIAN_FILTERING: Use the median filer of the solution in the post processing phase.
The first constructor initializes ``StereoVar`` with all the default parameters. So, you only have to set ``StereoVar::maxDisp`` and / or ``StereoVar::minDisp`` at minimum. The second constructor enables you to set each parameter to a custom value.
StereoVar::operator ()
-----------------------
.. ocv:function:: void StereoVar::operator()( const Mat& left, const Mat& right, Mat& disp )
Computes disparity using the variational algorithm for a rectified stereo pair.
:param left: Left 8-bit single-channel or 3-channel image.
:param right: Right image of the same size and the same type as the left one.
:param disp: Output disparity map. It is a 8-bit signed single-channel image of the same size as the input image.
The method executes the variational algorithm on a rectified stereo pair. See ``stereo_match.cpp`` OpenCV sample on how to prepare images and call the method.
**Note**:
The method is not constant, so you should not use the same ``StereoVar`` instance from different threads simultaneously.

638
modules/contrib/include/opencv2/contrib.hpp
View File

@@ -1,638 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#ifndef __OPENCV_CONTRIB_HPP__
#define __OPENCV_CONTRIB_HPP__
#include "opencv2/core.hpp"
#include "opencv2/imgproc.hpp"
#include "opencv2/features2d.hpp"
#include "opencv2/objdetect.hpp"
namespace cv
{
class CV_EXPORTS Octree
{
public:
struct Node
{
Node() { memset(this, 0, sizeof(Node)); }
int begin, end;
float x_min, x_max, y_min, y_max, z_min, z_max;
int maxLevels;
bool isLeaf;
int children[8];
};
Octree();
Octree( const std::vector<Point3f>& points, int maxLevels = 10, int minPoints = 20 );
virtual ~Octree();
virtual void buildTree( const std::vector<Point3f>& points, int maxLevels = 10, int minPoints = 20 );
virtual void getPointsWithinSphere( const Point3f& center, float radius,
std::vector<Point3f>& points ) const;
const std::vector<Node>& getNodes() const { return nodes; }
private:
int minPoints;
std::vector<Point3f> points;
std::vector<Node> nodes;
virtual void buildNext(size_t node_ind);
};
class CV_EXPORTS Mesh3D
{
public:
struct EmptyMeshException {};
Mesh3D();
Mesh3D(const std::vector<Point3f>& vtx);
~Mesh3D();
void buildOctree();
void clearOctree();
float estimateResolution(float tryRatio = 0.1f);
void computeNormals(float normalRadius, int minNeighbors = 20);
void computeNormals(const std::vector<int>& subset, float normalRadius, int minNeighbors = 20);
void writeAsVrml(const String& file, const std::vector<Scalar>& colors = std::vector<Scalar>()) const;
std::vector<Point3f> vtx;
std::vector<Point3f> normals;
float resolution;
Octree octree;
const static Point3f allzero;
};
class CV_EXPORTS SpinImageModel
{
public:
/* model parameters, leave unset for default or auto estimate */
float normalRadius;
int minNeighbors;
float binSize;
int imageWidth;
float lambda;
float gamma;
float T_GeometriccConsistency;
float T_GroupingCorespondances;
/* public interface */
SpinImageModel();
explicit SpinImageModel(const Mesh3D& mesh);
~SpinImageModel();
void selectRandomSubset(float ratio);
void setSubset(const std::vector<int>& subset);
void compute();
void match(const SpinImageModel& scene, std::vector< std::vector<Vec2i> >& result);
Mat packRandomScaledSpins(bool separateScale = false, size_t xCount = 10, size_t yCount = 10) const;
size_t getSpinCount() const { return spinImages.rows; }
Mat getSpinImage(size_t index) const { return spinImages.row((int)index); }
const Point3f& getSpinVertex(size_t index) const { return mesh.vtx[subset[index]]; }
const Point3f& getSpinNormal(size_t index) const { return mesh.normals[subset[index]]; }
const Mesh3D& getMesh() const { return mesh; }
Mesh3D& getMesh() { return mesh; }
/* static utility functions */
static bool spinCorrelation(const Mat& spin1, const Mat& spin2, float lambda, float& result);
static Point2f calcSpinMapCoo(const Point3f& point, const Point3f& vertex, const Point3f& normal);
static float geometricConsistency(const Point3f& pointScene1, const Point3f& normalScene1,
const Point3f& pointModel1, const Point3f& normalModel1,
const Point3f& pointScene2, const Point3f& normalScene2,
const Point3f& pointModel2, const Point3f& normalModel2);
static float groupingCreteria(const Point3f& pointScene1, const Point3f& normalScene1,
const Point3f& pointModel1, const Point3f& normalModel1,
const Point3f& pointScene2, const Point3f& normalScene2,
const Point3f& pointModel2, const Point3f& normalModel2,
float gamma);
protected:
void defaultParams();
void matchSpinToModel(const Mat& spin, std::vector<int>& indeces,
std::vector<float>& corrCoeffs, bool useExtremeOutliers = true) const;
void repackSpinImages(const std::vector<uchar>& mask, Mat& spinImages, bool reAlloc = true) const;
std::vector<int> subset;
Mesh3D mesh;
Mat spinImages;
};
class CV_EXPORTS TickMeter
{
public:
TickMeter();
void start();
void stop();
int64 getTimeTicks() const;
double getTimeMicro() const;
double getTimeMilli() const;
double getTimeSec() const;
int64 getCounter() const;
void reset();
private:
int64 counter;
int64 sumTime;
int64 startTime;
};
//CV_EXPORTS std::ostream& operator<<(std::ostream& out, const TickMeter& tm);
class CV_EXPORTS SelfSimDescriptor
{
public:
SelfSimDescriptor();
SelfSimDescriptor(int _ssize, int _lsize,
int _startDistanceBucket=DEFAULT_START_DISTANCE_BUCKET,
int _numberOfDistanceBuckets=DEFAULT_NUM_DISTANCE_BUCKETS,
int _nangles=DEFAULT_NUM_ANGLES);
SelfSimDescriptor(const SelfSimDescriptor& ss);
virtual ~SelfSimDescriptor();
SelfSimDescriptor& operator = (const SelfSimDescriptor& ss);
size_t getDescriptorSize() const;
Size getGridSize( Size imgsize, Size winStride ) const;
virtual void compute(const Mat& img, std::vector<float>& descriptors, Size winStride=Size(),
const std::vector<Point>& locations=std::vector<Point>()) const;
virtual void computeLogPolarMapping(Mat& mappingMask) const;
virtual void SSD(const Mat& img, Point pt, Mat& ssd) const;
int smallSize;
int largeSize;
int startDistanceBucket;
int numberOfDistanceBuckets;
int numberOfAngles;
enum { DEFAULT_SMALL_SIZE = 5, DEFAULT_LARGE_SIZE = 41,
DEFAULT_NUM_ANGLES = 20, DEFAULT_START_DISTANCE_BUCKET = 3,
DEFAULT_NUM_DISTANCE_BUCKETS = 7 };
};
CV_EXPORTS_W int chamerMatching( Mat& img, Mat& templ,
CV_OUT std::vector<std::vector<Point> >& results, CV_OUT std::vector<float>& cost,
double templScale=1, int maxMatches = 20,
double minMatchDistance = 1.0, int padX = 3,
int padY = 3, int scales = 5, double minScale = 0.6, double maxScale = 1.6,
double orientationWeight = 0.5, double truncate = 20);
class CV_EXPORTS_W StereoVar
{
public:
// Flags
enum {USE_INITIAL_DISPARITY = 1, USE_EQUALIZE_HIST = 2, USE_SMART_ID = 4, USE_AUTO_PARAMS = 8, USE_MEDIAN_FILTERING = 16};
enum {CYCLE_O, CYCLE_V};
enum {PENALIZATION_TICHONOV, PENALIZATION_CHARBONNIER, PENALIZATION_PERONA_MALIK};
//! the default constructor
CV_WRAP StereoVar();
//! the full constructor taking all the necessary algorithm parameters
CV_WRAP StereoVar(int levels, double pyrScale, int nIt, int minDisp, int maxDisp, int poly_n, double poly_sigma, float fi, float lambda, int penalization, int cycle, int flags);
//! the destructor
virtual ~StereoVar();
//! the stereo correspondence operator that computes disparity map for the specified rectified stereo pair
CV_WRAP_AS(compute) virtual void operator()(const Mat& left, const Mat& right, CV_OUT Mat& disp);
CV_PROP_RW int levels;
CV_PROP_RW double pyrScale;
CV_PROP_RW int nIt;
CV_PROP_RW int minDisp;
CV_PROP_RW int maxDisp;
CV_PROP_RW int poly_n;
CV_PROP_RW double poly_sigma;
CV_PROP_RW float fi;
CV_PROP_RW float lambda;
CV_PROP_RW int penalization;
CV_PROP_RW int cycle;
CV_PROP_RW int flags;
private:
void autoParams();
void FMG(Mat &I1, Mat &I2, Mat &I2x, Mat &u, int level);
void VCycle_MyFAS(Mat &I1_h, Mat &I2_h, Mat &I2x_h, Mat &u_h, int level);
void VariationalSolver(Mat &I1_h, Mat &I2_h, Mat &I2x_h, Mat &u_h, int level);
};
CV_EXPORTS void polyfit(const Mat& srcx, const Mat& srcy, Mat& dst, int order);
class CV_EXPORTS Directory
{
public:
static std::vector<String> GetListFiles ( const String& path, const String & exten = "*", bool addPath = true );
static std::vector<String> GetListFilesR ( const String& path, const String & exten = "*", bool addPath = true );
static std::vector<String> GetListFolders( const String& path, const String & exten = "*", bool addPath = true );
};
/*
* Generation of a set of different colors by the following way:
* 1) generate more then need colors (in "factor" times) in RGB,
* 2) convert them to Lab,
* 3) choose the needed count of colors from the set that are more different from
* each other,
* 4) convert the colors back to RGB
*/
CV_EXPORTS void generateColors( std::vector<Scalar>& colors, size_t count, size_t factor=100 );
/*
* Estimate the rigid body motion from frame0 to frame1. The method is based on the paper
* "Real-Time Visual Odometry from Dense RGB-D Images", F. Steinbucker, J. Strum, D. Cremers, ICCV, 2011.
*/
enum { ROTATION = 1,
TRANSLATION = 2,
RIGID_BODY_MOTION = 4
};
CV_EXPORTS bool RGBDOdometry( Mat& Rt, const Mat& initRt,
const Mat& image0, const Mat& depth0, const Mat& mask0,
const Mat& image1, const Mat& depth1, const Mat& mask1,
const Mat& cameraMatrix, float minDepth=0.f, float maxDepth=4.f, float maxDepthDiff=0.07f,
const std::vector<int>& iterCounts=std::vector<int>(),
const std::vector<float>& minGradientMagnitudes=std::vector<float>(),
int transformType=RIGID_BODY_MOTION );
/**
*Bilinear interpolation technique.
*
*The value of a desired cortical pixel is obtained through a bilinear interpolation of the values
*of the four nearest neighbouring Cartesian pixels to the center of the RF.
*The same principle is applied to the inverse transformation.
*
*More details can be found in http://dx.doi.org/10.1007/978-3-642-23968-7_5
*/
class CV_EXPORTS LogPolar_Interp
{
public:
LogPolar_Interp() {}
/**
*Constructor
*\param w the width of the input image
*\param h the height of the input image
*\param center the transformation center: where the output precision is maximal
*\param R the number of rings of the cortical image (default value 70 pixel)
*\param ro0 the radius of the blind spot (default value 3 pixel)
*\param full \a 1 (default value) means that the retinal image (the inverse transform) is computed within the circumscribing circle.
* \a 0 means that the retinal image is computed within the inscribed circle.
*\param S the number of sectors of the cortical image (default value 70 pixel).
* Its value is usually internally computed to obtain a pixel aspect ratio equals to 1.
*\param sp \a 1 (default value) means that the parameter \a S is internally computed.
* \a 0 means that the parameter \a S is provided by the user.
*/
LogPolar_Interp(int w, int h, Point2i center, int R=70, double ro0=3.0,
int interp=INTER_LINEAR, int full=1, int S=117, int sp=1);
/**
*Transformation from Cartesian image to cortical (log-polar) image.
*\param source the Cartesian image
*\return the transformed image (cortical image)
*/
const Mat to_cortical(const Mat &source);
/**
*Transformation from cortical image to retinal (inverse log-polar) image.
*\param source the cortical image
*\return the transformed image (retinal image)
*/
const Mat to_cartesian(const Mat &source);
/**
*Destructor
*/
~LogPolar_Interp();
protected:
Mat Rsri;
Mat Csri;
int S, R, M, N;
int top, bottom,left,right;
double ro0, romax, a, q;
int interp;
Mat ETAyx;
Mat CSIyx;
void create_map(int M, int N, int R, int S, double ro0);
};
/**
*Overlapping circular receptive fields technique
*
*The Cartesian plane is divided in two regions: the fovea and the periphery.
*The fovea (oversampling) is handled by using the bilinear interpolation technique described above, whereas in
*the periphery we use the overlapping Gaussian circular RFs.
*
*More details can be found in http://dx.doi.org/10.1007/978-3-642-23968-7_5
*/
class CV_EXPORTS LogPolar_Overlapping
{
public:
LogPolar_Overlapping() {}
/**
*Constructor
*\param w the width of the input image
*\param h the height of the input image
*\param center the transformation center: where the output precision is maximal
*\param R the number of rings of the cortical image (default value 70 pixel)
*\param ro0 the radius of the blind spot (default value 3 pixel)
*\param full \a 1 (default value) means that the retinal image (the inverse transform) is computed within the circumscribing circle.
* \a 0 means that the retinal image is computed within the inscribed circle.
*\param S the number of sectors of the cortical image (default value 70 pixel).
* Its value is usually internally computed to obtain a pixel aspect ratio equals to 1.
*\param sp \a 1 (default value) means that the parameter \a S is internally computed.
* \a 0 means that the parameter \a S is provided by the user.
*/
LogPolar_Overlapping(int w, int h, Point2i center, int R=70,
double ro0=3.0, int full=1, int S=117, int sp=1);
/**
*Transformation from Cartesian image to cortical (log-polar) image.
*\param source the Cartesian image
*\return the transformed image (cortical image)
*/
const Mat to_cortical(const Mat &source);
/**
*Transformation from cortical image to retinal (inverse log-polar) image.
*\param source the cortical image
*\return the transformed image (retinal image)
*/
const Mat to_cartesian(const Mat &source);
/**
*Destructor
*/
~LogPolar_Overlapping();
protected:
Mat Rsri;
Mat Csri;
std::vector<int> Rsr;
std::vector<int> Csr;
std::vector<double> Wsr;
int S, R, M, N, ind1;
int top, bottom,left,right;
double ro0, romax, a, q;
struct kernel
{
kernel() { w = 0; }
std::vector<double> weights;
int w;
};
Mat ETAyx;
Mat CSIyx;
std::vector<kernel> w_ker_2D;
void create_map(int M, int N, int R, int S, double ro0);
};
/**
* Adjacent receptive fields technique
*
*All the Cartesian pixels, whose coordinates in the cortical domain share the same integer part, are assigned to the same RF.
*The precision of the boundaries of the RF can be improved by breaking each pixel into subpixels and assigning each of them to the correct RF.
*This technique is implemented from: Traver, V., Pla, F.: Log-polar mapping template design: From task-level requirements
*to geometry parameters. Image Vision Comput. 26(10) (2008) 1354-1370
*
*More details can be found in http://dx.doi.org/10.1007/978-3-642-23968-7_5
*/
class CV_EXPORTS LogPolar_Adjacent
{
public:
LogPolar_Adjacent() {}
/**
*Constructor
*\param w the width of the input image
*\param h the height of the input image
*\param center the transformation center: where the output precision is maximal
*\param R the number of rings of the cortical image (default value 70 pixel)
*\param ro0 the radius of the blind spot (default value 3 pixel)
*\param smin the size of the subpixel (default value 0.25 pixel)
*\param full \a 1 (default value) means that the retinal image (the inverse transform) is computed within the circumscribing circle.
* \a 0 means that the retinal image is computed within the inscribed circle.
*\param S the number of sectors of the cortical image (default value 70 pixel).
* Its value is usually internally computed to obtain a pixel aspect ratio equals to 1.
*\param sp \a 1 (default value) means that the parameter \a S is internally computed.
* \a 0 means that the parameter \a S is provided by the user.
*/
LogPolar_Adjacent(int w, int h, Point2i center, int R=70, double ro0=3.0, double smin=0.25, int full=1, int S=117, int sp=1);
/**
*Transformation from Cartesian image to cortical (log-polar) image.
*\param source the Cartesian image
*\return the transformed image (cortical image)
*/
const Mat to_cortical(const Mat &source);
/**
*Transformation from cortical image to retinal (inverse log-polar) image.
*\param source the cortical image
*\return the transformed image (retinal image)
*/
const Mat to_cartesian(const Mat &source);
/**
*Destructor
*/
~LogPolar_Adjacent();
protected:
struct pixel
{
pixel() { u = v = 0; a = 0.; }
int u;
int v;
double a;
};
int S, R, M, N;
int top, bottom,left,right;
double ro0, romax, a, q;
std::vector<std::vector<pixel> > L;
std::vector<double> A;
void subdivide_recursively(double x, double y, int i, int j, double length, double smin);
bool get_uv(double x, double y, int&u, int&v);
void create_map(int M, int N, int R, int S, double ro0, double smin);
};
CV_EXPORTS Mat subspaceProject(InputArray W, InputArray mean, InputArray src);
CV_EXPORTS Mat subspaceReconstruct(InputArray W, InputArray mean, InputArray src);
class CV_EXPORTS LDA
{
public:
// Initializes a LDA with num_components (default 0) and specifies how
// samples are aligned (default dataAsRow=true).
LDA(int num_components = 0) :
_num_components(num_components) { }
// Initializes and performs a Discriminant Analysis with Fisher's
// Optimization Criterion on given data in src and corresponding labels
// in labels. If 0 (or less) number of components are given, they are
// automatically determined for given data in computation.
LDA(InputArrayOfArrays src, InputArray labels,
int num_components = 0) :
_num_components(num_components)
{
this->compute(src, labels); //! compute eigenvectors and eigenvalues
}
// Serializes this object to a given filename.
void save(const String& filename) const;
// Deserializes this object from a given filename.
void load(const String& filename);
// Serializes this object to a given cv::FileStorage.
void save(FileStorage& fs) const;
// Deserializes this object from a given cv::FileStorage.
void load(const FileStorage& node);
// Destructor.
~LDA() {}
//! Compute the discriminants for data in src and labels.
void compute(InputArrayOfArrays src, InputArray labels);
// Projects samples into the LDA subspace.
Mat project(InputArray src);
// Reconstructs projections from the LDA subspace.
Mat reconstruct(InputArray src);
// Returns the eigenvectors of this LDA.
Mat eigenvectors() const { return _eigenvectors; }
// Returns the eigenvalues of this LDA.
Mat eigenvalues() const { return _eigenvalues; }
protected:
bool _dataAsRow;
int _num_components;
Mat _eigenvectors;
Mat _eigenvalues;
void lda(InputArrayOfArrays src, InputArray labels);
};
class CV_EXPORTS_W FaceRecognizer : public Algorithm
{
public:
//! virtual destructor
virtual ~FaceRecognizer() {}
// Trains a FaceRecognizer.
CV_WRAP virtual void train(InputArrayOfArrays src, InputArray labels) = 0;
// Updates a FaceRecognizer.
CV_WRAP virtual void update(InputArrayOfArrays src, InputArray labels);
// Gets a prediction from a FaceRecognizer.
virtual int predict(InputArray src) const = 0;
// Predicts the label and confidence for a given sample.
CV_WRAP virtual void predict(InputArray src, CV_OUT int &label, CV_OUT double &confidence) const = 0;
// Serializes this object to a given filename.
CV_WRAP virtual void save(const String& filename) const;
// Deserializes this object from a given filename.
CV_WRAP virtual void load(const String& filename);
// Serializes this object to a given cv::FileStorage.
virtual void save(FileStorage& fs) const = 0;
// Deserializes this object from a given cv::FileStorage.
virtual void load(const FileStorage& fs) = 0;
};
CV_EXPORTS_W Ptr<FaceRecognizer> createEigenFaceRecognizer(int num_components = 0, double threshold = DBL_MAX);
CV_EXPORTS_W Ptr<FaceRecognizer> createFisherFaceRecognizer(int num_components = 0, double threshold = DBL_MAX);
CV_EXPORTS_W Ptr<FaceRecognizer> createLBPHFaceRecognizer(int radius=1, int neighbors=8,
int grid_x=8, int grid_y=8, double threshold = DBL_MAX);
enum
{
COLORMAP_AUTUMN = 0,
COLORMAP_BONE = 1,
COLORMAP_JET = 2,
COLORMAP_WINTER = 3,
COLORMAP_RAINBOW = 4,
COLORMAP_OCEAN = 5,
COLORMAP_SUMMER = 6,
COLORMAP_SPRING = 7,
COLORMAP_COOL = 8,
COLORMAP_HSV = 9,
COLORMAP_PINK = 10,
COLORMAP_HOT = 11
};
CV_EXPORTS_W void applyColorMap(InputArray src, OutputArray dst, int colormap);
CV_EXPORTS bool initModule_contrib();
}
#include "opencv2/contrib/openfabmap.hpp"
#endif

384
modules/contrib/include/opencv2/contrib/compat.hpp
View File

@@ -1,384 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2009-2011, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#ifndef __OPENCV_CONTRIB_COMPAT_HPP__
#define __OPENCV_CONTRIB_COMPAT_HPP__
#include "opencv2/core/core_c.h"
#ifdef __cplusplus
/****************************************************************************************\
* Adaptive Skin Detector *
\****************************************************************************************/
class CV_EXPORTS CvAdaptiveSkinDetector
{
private:
enum {
GSD_HUE_LT = 3,
GSD_HUE_UT = 33,
GSD_INTENSITY_LT = 15,
GSD_INTENSITY_UT = 250
};
class CV_EXPORTS Histogram
{
private:
enum {
HistogramSize = (GSD_HUE_UT - GSD_HUE_LT + 1)
};
protected:
int findCoverageIndex(double surfaceToCover, int defaultValue = 0);
public:
CvHistogram *fHistogram;
Histogram();
virtual ~Histogram();
void findCurveThresholds(int &x1, int &x2, double percent = 0.05);
void mergeWith(Histogram *source, double weight);
};
int nStartCounter, nFrameCount, nSkinHueLowerBound, nSkinHueUpperBound, nMorphingMethod, nSamplingDivider;
double fHistogramMergeFactor, fHuePercentCovered;
Histogram histogramHueMotion, skinHueHistogram;
IplImage *imgHueFrame, *imgSaturationFrame, *imgLastGrayFrame, *imgMotionFrame, *imgFilteredFrame;
IplImage *imgShrinked, *imgTemp, *imgGrayFrame, *imgHSVFrame;
protected:
void initData(IplImage *src, int widthDivider, int heightDivider);
void adaptiveFilter();
public:
enum {
MORPHING_METHOD_NONE = 0,
MORPHING_METHOD_ERODE = 1,
MORPHING_METHOD_ERODE_ERODE = 2,
MORPHING_METHOD_ERODE_DILATE = 3
};
CvAdaptiveSkinDetector(int samplingDivider = 1, int morphingMethod = MORPHING_METHOD_NONE);
virtual ~CvAdaptiveSkinDetector();
virtual void process(IplImage *inputBGRImage, IplImage *outputHueMask);
};
/****************************************************************************************\
* Fuzzy MeanShift Tracker *
\****************************************************************************************/
class CV_EXPORTS CvFuzzyPoint {
public:
double x, y, value;
CvFuzzyPoint(double _x, double _y);
};
class CV_EXPORTS CvFuzzyCurve {
private:
std::vector<CvFuzzyPoint> points;
double value, centre;
bool between(double x, double x1, double x2);
public:
CvFuzzyCurve();
~CvFuzzyCurve();
void setCentre(double _centre);
double getCentre();
void clear();
void addPoint(double x, double y);
double calcValue(double param);
double getValue();
void setValue(double _value);
};
class CV_EXPORTS CvFuzzyFunction {
public:
std::vector<CvFuzzyCurve> curves;
CvFuzzyFunction();
~CvFuzzyFunction();
void addCurve(CvFuzzyCurve *curve, double value = 0);
void resetValues();
double calcValue();
CvFuzzyCurve *newCurve();
};
class CV_EXPORTS CvFuzzyRule {
private:
CvFuzzyCurve *fuzzyInput1, *fuzzyInput2;
CvFuzzyCurve *fuzzyOutput;
public:
CvFuzzyRule();
~CvFuzzyRule();
void setRule(CvFuzzyCurve *c1, CvFuzzyCurve *c2, CvFuzzyCurve *o1);
double calcValue(double param1, double param2);
CvFuzzyCurve *getOutputCurve();
};
class CV_EXPORTS CvFuzzyController {
private:
std::vector<CvFuzzyRule*> rules;
public:
CvFuzzyController();
~CvFuzzyController();
void addRule(CvFuzzyCurve *c1, CvFuzzyCurve *c2, CvFuzzyCurve *o1);
double calcOutput(double param1, double param2);
};
class CV_EXPORTS CvFuzzyMeanShiftTracker
{
private:
class FuzzyResizer
{
private:
CvFuzzyFunction iInput, iOutput;
CvFuzzyController fuzzyController;
public:
FuzzyResizer();
int calcOutput(double edgeDensity, double density);
};
class SearchWindow
{
public:
FuzzyResizer *fuzzyResizer;
int x, y;
int width, height, maxWidth, maxHeight, ellipseHeight, ellipseWidth;
int ldx, ldy, ldw, ldh, numShifts, numIters;
int xGc, yGc;
long m00, m01, m10, m11, m02, m20;
double ellipseAngle;
double density;
unsigned int depthLow, depthHigh;
int verticalEdgeLeft, verticalEdgeRight, horizontalEdgeTop, horizontalEdgeBottom;
SearchWindow();
~SearchWindow();
void setSize(int _x, int _y, int _width, int _height);
void initDepthValues(IplImage *maskImage, IplImage *depthMap);
bool shift();
void extractInfo(IplImage *maskImage, IplImage *depthMap, bool initDepth);
void getResizeAttribsEdgeDensityLinear(int &resizeDx, int &resizeDy, int &resizeDw, int &resizeDh);
void getResizeAttribsInnerDensity(int &resizeDx, int &resizeDy, int &resizeDw, int &resizeDh);
void getResizeAttribsEdgeDensityFuzzy(int &resizeDx, int &resizeDy, int &resizeDw, int &resizeDh);
bool meanShift(IplImage *maskImage, IplImage *depthMap, int maxIteration, bool initDepth);
};
public:
enum TrackingState
{
tsNone = 0,
tsSearching = 1,
tsTracking = 2,
tsSetWindow = 3,
tsDisabled = 10
};
enum ResizeMethod {
rmEdgeDensityLinear = 0,
rmEdgeDensityFuzzy = 1,
rmInnerDensity = 2
};
enum {
MinKernelMass = 1000
};
SearchWindow kernel;
int searchMode;
private:
enum
{
MaxMeanShiftIteration = 5,
MaxSetSizeIteration = 5
};
void findOptimumSearchWindow(SearchWindow &searchWindow, IplImage *maskImage, IplImage *depthMap, int maxIteration, int resizeMethod, bool initDepth);
public:
CvFuzzyMeanShiftTracker();
~CvFuzzyMeanShiftTracker();
void track(IplImage *maskImage, IplImage *depthMap, int resizeMethod, bool resetSearch, int minKernelMass = MinKernelMass);
};
namespace cv
{
typedef bool (*BundleAdjustCallback)(int iteration, double norm_error, void* user_data);
class CV_EXPORTS LevMarqSparse {
public:
LevMarqSparse();
LevMarqSparse(int npoints, // number of points
int ncameras, // number of cameras
int nPointParams, // number of params per one point (3 in case of 3D points)
int nCameraParams, // number of parameters per one camera
int nErrParams, // number of parameters in measurement vector
// for 1 point at one camera (2 in case of 2D projections)
Mat& visibility, // visibility matrix. rows correspond to points, columns correspond to cameras
// 1 - point is visible for the camera, 0 - invisible
Mat& P0, // starting vector of parameters, first cameras then points
Mat& X, // measurements, in order of visibility. non visible cases are skipped
TermCriteria criteria, // termination criteria
// callback for estimation of Jacobian matrices
void (*fjac)(int i, int j, Mat& point_params,
Mat& cam_params, Mat& A, Mat& B, void* data),
// callback for estimation of backprojection errors
void (*func)(int i, int j, Mat& point_params,
Mat& cam_params, Mat& estim, void* data),
void* data, // user-specific data passed to the callbacks
BundleAdjustCallback cb, void* user_data
);
virtual ~LevMarqSparse();
virtual void run( int npoints, // number of points
int ncameras, // number of cameras
int nPointParams, // number of params per one point (3 in case of 3D points)
int nCameraParams, // number of parameters per one camera
int nErrParams, // number of parameters in measurement vector
// for 1 point at one camera (2 in case of 2D projections)
Mat& visibility, // visibility matrix. rows correspond to points, columns correspond to cameras
// 1 - point is visible for the camera, 0 - invisible
Mat& P0, // starting vector of parameters, first cameras then points
Mat& X, // measurements, in order of visibility. non visible cases are skipped
TermCriteria criteria, // termination criteria
// callback for estimation of Jacobian matrices
void (CV_CDECL * fjac)(int i, int j, Mat& point_params,
Mat& cam_params, Mat& A, Mat& B, void* data),
// callback for estimation of backprojection errors
void (CV_CDECL * func)(int i, int j, Mat& point_params,
Mat& cam_params, Mat& estim, void* data),
void* data // user-specific data passed to the callbacks
);
virtual void clear();
// useful function to do simple bundle adjustment tasks
static void bundleAdjust(std::vector<Point3d>& points, // positions of points in global coordinate system (input and output)
const std::vector<std::vector<Point2d> >& imagePoints, // projections of 3d points for every camera
const std::vector<std::vector<int> >& visibility, // visibility of 3d points for every camera
std::vector<Mat>& cameraMatrix, // intrinsic matrices of all cameras (input and output)
std::vector<Mat>& R, // rotation matrices of all cameras (input and output)
std::vector<Mat>& T, // translation vector of all cameras (input and output)
std::vector<Mat>& distCoeffs, // distortion coefficients of all cameras (input and output)
const TermCriteria& criteria=
TermCriteria(TermCriteria::COUNT+TermCriteria::EPS, 30, DBL_EPSILON),
BundleAdjustCallback cb = 0, void* user_data = 0);
public:
virtual void optimize(CvMat &_vis); //main function that runs minimization
//iteratively asks for measurement for visible camera-point pairs
void ask_for_proj(CvMat &_vis,bool once=false);
//iteratively asks for Jacobians for every camera_point pair
void ask_for_projac(CvMat &_vis);
CvMat* err; //error X-hX
double prevErrNorm, errNorm;
double lambda;
CvTermCriteria criteria;
int iters;
CvMat** U; //size of array is equal to number of cameras
CvMat** V; //size of array is equal to number of points
CvMat** inv_V_star; //inverse of V*
CvMat** A;
CvMat** B;
CvMat** W;
CvMat* X; //measurement
CvMat* hX; //current measurement extimation given new parameter vector
CvMat* prevP; //current already accepted parameter.
CvMat* P; // parameters used to evaluate function with new params
// this parameters may be rejected
CvMat* deltaP; //computed increase of parameters (result of normal system solution )
CvMat** ea; // sum_i AijT * e_ij , used as right part of normal equation
// length of array is j = number of cameras
CvMat** eb; // sum_j BijT * e_ij , used as right part of normal equation
// length of array is i = number of points
CvMat** Yj; //length of array is i = num_points
CvMat* S; //big matrix of block Sjk , each block has size num_cam_params x num_cam_params
CvMat* JtJ_diag; //diagonal of JtJ, used to backup diagonal elements before augmentation
CvMat* Vis_index; // matrix which element is index of measurement for point i and camera j
int num_cams;
int num_points;
int num_err_param;
int num_cam_param;
int num_point_param;
//target function and jacobian pointers, which needs to be initialized
void (*fjac)(int i, int j, Mat& point_params, Mat& cam_params, Mat& A, Mat& B, void* data);
void (*func)(int i, int j, Mat& point_params, Mat& cam_params, Mat& estim, void* data);
void* data;
BundleAdjustCallback cb;
void* user_data;
};
} // cv
#endif /* __cplusplus */
#endif /* __OPENCV_CONTRIB_COMPAT_HPP__ */

219
modules/contrib/include/opencv2/contrib/hybridtracker.hpp
View File

@@ -1,219 +0,0 @@
//*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2008-2011, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of Intel Corporation may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#ifndef __OPENCV_HYBRIDTRACKER_H_
#define __OPENCV_HYBRIDTRACKER_H_
#include "opencv2/core.hpp"
#include "opencv2/imgproc.hpp"
#include "opencv2/features2d.hpp"
#include "opencv2/video/tracking.hpp"
#include "opencv2/ml.hpp"
#ifdef __cplusplus
namespace cv
{
// Motion model for tracking algorithm. Currently supports objects that do not move much.
// To add Kalman filter
struct CV_EXPORTS CvMotionModel
{
enum {LOW_PASS_FILTER = 0, KALMAN_FILTER = 1, EM = 2};
CvMotionModel()
{
}
float low_pass_gain; // low pass gain
};
// Mean Shift Tracker parameters for specifying use of HSV channel and CamShift parameters.
struct CV_EXPORTS CvMeanShiftTrackerParams
{
enum { H = 0, HS = 1, HSV = 2 };
CvMeanShiftTrackerParams(int tracking_type = CvMeanShiftTrackerParams::HS,
CvTermCriteria term_crit = CvTermCriteria());
int tracking_type;
std::vector<float> h_range;
std::vector<float> s_range;
std::vector<float> v_range;
CvTermCriteria term_crit;
};
// Feature tracking parameters
struct CV_EXPORTS CvFeatureTrackerParams
{
enum { SIFT = 0, SURF = 1, OPTICAL_FLOW = 2 };
CvFeatureTrackerParams(int featureType = 0, int windowSize = 0)
{
feature_type = featureType;
window_size = windowSize;
}
int feature_type; // Feature type to use
int window_size; // Window size in pixels around which to search for new window
};
// Hybrid Tracking parameters for specifying weights of individual trackers and motion model.
struct CV_EXPORTS CvHybridTrackerParams
{
CvHybridTrackerParams(float ft_tracker_weight = 0.5, float ms_tracker_weight = 0.5,
CvFeatureTrackerParams ft_params = CvFeatureTrackerParams(),
CvMeanShiftTrackerParams ms_params = CvMeanShiftTrackerParams(),
CvMotionModel model = CvMotionModel());
float ft_tracker_weight;
float ms_tracker_weight;
CvFeatureTrackerParams ft_params;
CvMeanShiftTrackerParams ms_params;
int motion_model;
float low_pass_gain;
};
// Performs Camshift using parameters from MeanShiftTrackerParams
class CV_EXPORTS CvMeanShiftTracker
{
private:
Mat hsv, hue;
Mat backproj;
Mat mask, maskroi;
MatND hist;
Rect prev_trackwindow;
RotatedRect prev_trackbox;
Point2f prev_center;
public:
CvMeanShiftTrackerParams params;
CvMeanShiftTracker();
explicit CvMeanShiftTracker(CvMeanShiftTrackerParams _params);
~CvMeanShiftTracker();
void newTrackingWindow(Mat image, Rect selection);
RotatedRect updateTrackingWindow(Mat image);
Mat getHistogramProjection(int type);
void setTrackingWindow(Rect _window);
Rect getTrackingWindow();
RotatedRect getTrackingEllipse();
Point2f getTrackingCenter();
};
// Performs SIFT/SURF feature tracking using parameters from FeatureTrackerParams
class CV_EXPORTS CvFeatureTracker
{
private:
Ptr<Feature2D> dd;
Ptr<DescriptorMatcher> matcher;
std::vector<DMatch> matches;
Mat prev_image;
Mat prev_image_bw;
Rect prev_trackwindow;
Point2d prev_center;
int ittr;
std::vector<Point2f> features[2];
public:
Mat disp_matches;
CvFeatureTrackerParams params;
CvFeatureTracker();
explicit CvFeatureTracker(CvFeatureTrackerParams params);
~CvFeatureTracker();
void newTrackingWindow(Mat image, Rect selection);
Rect updateTrackingWindow(Mat image);
Rect updateTrackingWindowWithSIFT(Mat image);
Rect updateTrackingWindowWithFlow(Mat image);
void setTrackingWindow(Rect _window);
Rect getTrackingWindow();
Point2f getTrackingCenter();
};
// Performs Hybrid Tracking and combines individual trackers using EM or filters
class CV_EXPORTS CvHybridTracker
{
private:
CvMeanShiftTracker* mstracker;
CvFeatureTracker* fttracker;
CvMat* samples;
CvMat* labels;
Rect prev_window;
Point2f prev_center;
Mat prev_proj;
RotatedRect trackbox;
int ittr;
Point2f curr_center;
inline float getL2Norm(Point2f p1, Point2f p2);
Mat getDistanceProjection(Mat image, Point2f center);
Mat getGaussianProjection(Mat image, int ksize, double sigma, Point2f center);
void updateTrackerWithEM(Mat image);
void updateTrackerWithLowPassFilter(Mat image);
public:
CvHybridTrackerParams params;
CvHybridTracker();
explicit CvHybridTracker(CvHybridTrackerParams params);
~CvHybridTracker();
void newTracker(Mat image, Rect selection);
void updateTracker(Mat image);
Rect getTrackingWindow();
};
typedef CvMotionModel MotionModel;
typedef CvMeanShiftTrackerParams MeanShiftTrackerParams;
typedef CvFeatureTrackerParams FeatureTrackerParams;
typedef CvHybridTrackerParams HybridTrackerParams;
typedef CvMeanShiftTracker MeanShiftTracker;
typedef CvFeatureTracker FeatureTracker;
typedef CvHybridTracker HybridTracker;
}
#endif
#endif

401
modules/contrib/include/opencv2/contrib/openfabmap.hpp
View File

@@ -1,401 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
// This file originates from the openFABMAP project:
// [http://code.google.com/p/openfabmap/]
//
// For published work which uses all or part of OpenFABMAP, please cite:
// [http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6224843]
//
// Original Algorithm by Mark Cummins and Paul Newman:
// [http://ijr.sagepub.com/content/27/6/647.short]
// [http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5613942]
// [http://ijr.sagepub.com/content/30/9/1100.abstract]
//
// License Agreement
//
// Copyright (C) 2012 Arren Glover [aj.glover@qut.edu.au] and
// Will Maddern [w.maddern@qut.edu.au], all rights reserved.
//
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#ifndef __OPENCV_OPENFABMAP_H_
#define __OPENCV_OPENFABMAP_H_
#include "opencv2/core.hpp"
#include "opencv2/features2d.hpp"
#include <vector>
#include <list>
#include <map>
#include <set>
#include <valarray>
namespace cv {
namespace of2 {
/*
Return data format of a FABMAP compare call
*/
struct CV_EXPORTS IMatch {
IMatch() :
queryIdx(-1), imgIdx(-1), likelihood(-DBL_MAX), match(-DBL_MAX) {
}
IMatch(int _queryIdx, int _imgIdx, double _likelihood, double _match) :
queryIdx(_queryIdx), imgIdx(_imgIdx), likelihood(_likelihood), match(
_match) {
}
int queryIdx; //query index
int imgIdx; //test index
double likelihood; //raw loglikelihood
double match; //normalised probability
bool operator<(const IMatch& m) const {
return match < m.match;
}
};
/*
Base FabMap class. Each FabMap method inherits from this class.
*/
class CV_EXPORTS FabMap {
public:
//FabMap options
enum {
MEAN_FIELD = 1,
SAMPLED = 2,
NAIVE_BAYES = 4,
CHOW_LIU = 8,
MOTION_MODEL = 16
};
FabMap(const Mat& clTree, double PzGe, double PzGNe, int flags,
int numSamples = 0);
virtual ~FabMap();
//methods to add training data for sampling method
virtual void addTraining(const Mat& queryImgDescriptor);
virtual void addTraining(const std::vector<Mat>& queryImgDescriptors);
//methods to add to the test data
virtual void add(const Mat& queryImgDescriptor);
virtual void add(const std::vector<Mat>& queryImgDescriptors);
//accessors
const std::vector<Mat>& getTrainingImgDescriptors() const;
const std::vector<Mat>& getTestImgDescriptors() const;
//Main FabMap image comparison
void compare(const Mat& queryImgDescriptor,
std::vector<IMatch>& matches, bool addQuery = false,
const Mat& mask = Mat());
void compare(const Mat& queryImgDescriptor,
const Mat& testImgDescriptors, std::vector<IMatch>& matches,
const Mat& mask = Mat());
void compare(const Mat& queryImgDescriptor,
const std::vector<Mat>& testImgDescriptors,
std::vector<IMatch>& matches, const Mat& mask = Mat());
void compare(const std::vector<Mat>& queryImgDescriptors, std::vector<
IMatch>& matches, bool addQuery = false, const Mat& mask =
Mat());
void compare(const std::vector<Mat>& queryImgDescriptors,
const std::vector<Mat>& testImgDescriptors,
std::vector<IMatch>& matches, const Mat& mask = Mat());
protected:
void compareImgDescriptor(const Mat& queryImgDescriptor,
int queryIndex, const std::vector<Mat>& testImgDescriptors,
std::vector<IMatch>& matches);
void addImgDescriptor(const Mat& queryImgDescriptor);
//the getLikelihoods method is overwritten for each different FabMap
//method.
virtual void getLikelihoods(const Mat& queryImgDescriptor,
const std::vector<Mat>& testImgDescriptors,
std::vector<IMatch>& matches);
virtual double getNewPlaceLikelihood(const Mat& queryImgDescriptor);
//turn likelihoods into probabilities (also add in motion model if used)
void normaliseDistribution(std::vector<IMatch>& matches);
//Chow-Liu Tree
int pq(int q);
double Pzq(int q, bool zq);
double PzqGzpq(int q, bool zq, bool zpq);
//FAB-MAP Core
double PzqGeq(bool zq, bool eq);
double PeqGL(int q, bool Lzq, bool eq);
double PzqGL(int q, bool zq, bool zpq, bool Lzq);
double PzqGzpqL(int q, bool zq, bool zpq, bool Lzq);
double (FabMap::*PzGL)(int q, bool zq, bool zpq, bool Lzq);
//data
Mat clTree;
std::vector<Mat> trainingImgDescriptors;
std::vector<Mat> testImgDescriptors;
std::vector<IMatch> priorMatches;
//parameters
double PzGe;
double PzGNe;
double Pnew;
double mBias;
double sFactor;
int flags;
int numSamples;
};
/*
The original FAB-MAP algorithm, developed based on:
http://ijr.sagepub.com/content/27/6/647.short
*/
class CV_EXPORTS FabMap1: public FabMap {
public:
FabMap1(const Mat& clTree, double PzGe, double PzGNe, int flags,
int numSamples = 0);
virtual ~FabMap1();
protected:
//FabMap1 implementation of likelihood comparison
void getLikelihoods(const Mat& queryImgDescriptor, const std::vector<
Mat>& testImgDescriptors, std::vector<IMatch>& matches);
};
/*
A computationally faster version of the original FAB-MAP algorithm. A look-
up-table is used to precompute many of the reoccuring calculations
*/
class CV_EXPORTS FabMapLUT: public FabMap {
public:
FabMapLUT(const Mat& clTree, double PzGe, double PzGNe,
int flags, int numSamples = 0, int precision = 6);
virtual ~FabMapLUT();
protected:
//FabMap look-up-table implementation of the likelihood comparison
void getLikelihoods(const Mat& queryImgDescriptor, const std::vector<
Mat>& testImgDescriptors, std::vector<IMatch>& matches);
//precomputed data
int (*table)[8];
//data precision
int precision;
};
/*
The Accelerated FAB-MAP algorithm, developed based on:
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5613942
*/
class CV_EXPORTS FabMapFBO: public FabMap {
public:
FabMapFBO(const Mat& clTree, double PzGe, double PzGNe, int flags,
int numSamples = 0, double rejectionThreshold = 1e-8, double PsGd =
1e-8, int bisectionStart = 512, int bisectionIts = 9);
virtual ~FabMapFBO();
protected:
//FabMap Fast Bail-out implementation of the likelihood comparison
void getLikelihoods(const Mat& queryImgDescriptor, const std::vector<
Mat>& testImgDescriptors, std::vector<IMatch>& matches);
//stucture used to determine word comparison order
struct WordStats {
WordStats() :
q(0), info(0), V(0), M(0) {
}
WordStats(int _q, double _info) :
q(_q), info(_info), V(0), M(0) {
}
int q;
double info;
mutable double V;
mutable double M;
bool operator<(const WordStats& w) const {
return info < w.info;
}
};
//private fast bail-out necessary functions
void setWordStatistics(const Mat& queryImgDescriptor, std::multiset<WordStats>& wordData);
double limitbisection(double v, double m);
double bennettInequality(double v, double m, double delta);
static bool compInfo(const WordStats& first, const WordStats& second);
//parameters
double PsGd;
double rejectionThreshold;
int bisectionStart;
int bisectionIts;
};
/*
The FAB-MAP2.0 algorithm, developed based on:
http://ijr.sagepub.com/content/30/9/1100.abstract
*/
class CV_EXPORTS FabMap2: public FabMap {
public:
FabMap2(const Mat& clTree, double PzGe, double PzGNe, int flags);
virtual ~FabMap2();
//FabMap2 builds the inverted index and requires an additional training/test
//add function
void addTraining(const Mat& queryImgDescriptors) {
FabMap::addTraining(queryImgDescriptors);
}
void addTraining(const std::vector<Mat>& queryImgDescriptors);
void add(const Mat& queryImgDescriptors) {
FabMap::add(queryImgDescriptors);
}
void add(const std::vector<Mat>& queryImgDescriptors);
protected:
//FabMap2 implementation of the likelihood comparison
void getLikelihoods(const Mat& queryImgDescriptor, const std::vector<
Mat>& testImgDescriptors, std::vector<IMatch>& matches);
double getNewPlaceLikelihood(const Mat& queryImgDescriptor);
//the likelihood function using the inverted index
void getIndexLikelihoods(const Mat& queryImgDescriptor, std::vector<
double>& defaults, std::map<int, std::vector<int> >& invertedMap,
std::vector<IMatch>& matches);
void addToIndex(const Mat& queryImgDescriptor,
std::vector<double>& defaults,
std::map<int, std::vector<int> >& invertedMap);
//data
std::vector<double> d1, d2, d3, d4;
std::vector<std::vector<int> > children;
// TODO: inverted map a vector?
std::vector<double> trainingDefaults;
std::map<int, std::vector<int> > trainingInvertedMap;
std::vector<double> testDefaults;
std::map<int, std::vector<int> > testInvertedMap;
};
/*
A Chow-Liu tree is required by FAB-MAP. The Chow-Liu tree provides an
estimate of the full distribution of visual words using a minimum spanning
tree. The tree is generated through training data.
*/
class CV_EXPORTS ChowLiuTree {
public:
ChowLiuTree();
virtual ~ChowLiuTree();
//add data to the chow-liu tree before calling make
void add(const Mat& imgDescriptor);
void add(const std::vector<Mat>& imgDescriptors);
const std::vector<Mat>& getImgDescriptors() const;
Mat make(double infoThreshold = 0.0);
private:
std::vector<Mat> imgDescriptors;
Mat mergedImgDescriptors;
typedef struct info {
float score;
short word1;
short word2;
} info;
//probabilities extracted from mergedImgDescriptors
double P(int a, bool za);
double JP(int a, bool za, int b, bool zb); //a & b
double CP(int a, bool za, int b, bool zb); // a | b
//calculating mutual information of all edges
void createBaseEdges(std::list<info>& edges, double infoThreshold);
double calcMutInfo(int word1, int word2);
static bool sortInfoScores(const info& first, const info& second);
//selecting minimum spanning egdges with maximum information
bool reduceEdgesToMinSpan(std::list<info>& edges);
//building the tree sctructure
Mat buildTree(int root_word, std::list<info> &edges);
void recAddToTree(Mat &cltree, int q, int pq,
std::list<info> &remaining_edges);
std::vector<int> extractChildren(std::list<info> &remaining_edges, int q);
};
/*
A custom vocabulary training method based on:
http://www.springerlink.com/content/d1h6j8x552532003/
*/
class CV_EXPORTS BOWMSCTrainer: public BOWTrainer {
public:
BOWMSCTrainer(double clusterSize = 0.4);
virtual ~BOWMSCTrainer();
// Returns trained vocabulary (i.e. cluster centers).
virtual Mat cluster() const;
virtual Mat cluster(const Mat& descriptors) const;
protected:
double clusterSize;
};
}
}
#endif /* OPENFABMAP_H_ */

288
modules/contrib/src/adaptiveskindetector.cpp
View File

@@ -1,288 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install, copy or use the software.
//
// Copyright (C) 2009, Farhad Dadgostar
// Intel Corporation and third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of Intel Corporation may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "precomp.hpp"
#include "opencv2/imgproc/imgproc_c.h"
#include "opencv2/contrib/compat.hpp"
#define ASD_INTENSITY_SET_PIXEL(pointer, qq) {(*pointer) = (unsigned char)qq;}
#define ASD_IS_IN_MOTION(pointer, v, threshold) ((abs((*(pointer)) - (v)) > (threshold)) ? true : false)
void CvAdaptiveSkinDetector::initData(IplImage *src, int widthDivider, int heightDivider)
{
CvSize imageSize = cvSize(src->width/widthDivider, src->height/heightDivider);
imgHueFrame = cvCreateImage(imageSize, IPL_DEPTH_8U, 1);
imgShrinked = cvCreateImage(imageSize, IPL_DEPTH_8U, src->nChannels);
imgSaturationFrame = cvCreateImage(imageSize, IPL_DEPTH_8U, 1);
imgMotionFrame = cvCreateImage(imageSize, IPL_DEPTH_8U, 1);
imgTemp = cvCreateImage(imageSize, IPL_DEPTH_8U, 1);
imgFilteredFrame = cvCreateImage(imageSize, IPL_DEPTH_8U, 1);
imgGrayFrame = cvCreateImage(imageSize, IPL_DEPTH_8U, 1);
imgLastGrayFrame = cvCreateImage(imageSize, IPL_DEPTH_8U, 1);
imgHSVFrame = cvCreateImage(imageSize, IPL_DEPTH_8U, 3);
}
CvAdaptiveSkinDetector::CvAdaptiveSkinDetector(int samplingDivider, int morphingMethod)
{
nSkinHueLowerBound = GSD_HUE_LT;
nSkinHueUpperBound = GSD_HUE_UT;
fHistogramMergeFactor = 0.05; // empirical result
fHuePercentCovered = 0.95; // empirical result
nMorphingMethod = morphingMethod;
nSamplingDivider = samplingDivider;
nFrameCount = 0;
nStartCounter = 0;
imgHueFrame = NULL;
imgMotionFrame = NULL;
imgTemp = NULL;
imgFilteredFrame = NULL;
imgShrinked = NULL;
imgGrayFrame = NULL;
imgLastGrayFrame = NULL;
imgSaturationFrame = NULL;
imgHSVFrame = NULL;
}
CvAdaptiveSkinDetector::~CvAdaptiveSkinDetector()
{
cvReleaseImage(&imgHueFrame);
cvReleaseImage(&imgSaturationFrame);
cvReleaseImage(&imgMotionFrame);
cvReleaseImage(&imgTemp);
cvReleaseImage(&imgFilteredFrame);
cvReleaseImage(&imgShrinked);
cvReleaseImage(&imgGrayFrame);
cvReleaseImage(&imgLastGrayFrame);
cvReleaseImage(&imgHSVFrame);
}
void CvAdaptiveSkinDetector::process(IplImage *inputBGRImage, IplImage *outputHueMask)
{
IplImage *src = inputBGRImage;
int h, v, i, l;
bool isInit = false;
nFrameCount++;
if (imgHueFrame == NULL)
{
isInit = true;
initData(src, nSamplingDivider, nSamplingDivider);
}
unsigned char *pShrinked, *pHueFrame, *pMotionFrame, *pLastGrayFrame, *pFilteredFrame, *pGrayFrame;
pShrinked = (unsigned char *)imgShrinked->imageData;
pHueFrame = (unsigned char *)imgHueFrame->imageData;
pMotionFrame = (unsigned char *)imgMotionFrame->imageData;
pLastGrayFrame = (unsigned char *)imgLastGrayFrame->imageData;
pFilteredFrame = (unsigned char *)imgFilteredFrame->imageData;
pGrayFrame = (unsigned char *)imgGrayFrame->imageData;
if ((src->width != imgHueFrame->width) || (src->height != imgHueFrame->height))
{
cvResize(src, imgShrinked);
cvCvtColor(imgShrinked, imgHSVFrame, CV_BGR2HSV);
}
else
{
cvCvtColor(src, imgHSVFrame, CV_BGR2HSV);
}
cvSplit(imgHSVFrame, imgHueFrame, imgSaturationFrame, imgGrayFrame, 0);
cvSetZero(imgMotionFrame);
cvSetZero(imgFilteredFrame);
l = imgHueFrame->height * imgHueFrame->width;
for (i = 0; i < l; i++)
{
v = (*pGrayFrame);
if ((v >= GSD_INTENSITY_LT) && (v <= GSD_INTENSITY_UT))
{
h = (*pHueFrame);
if ((h >= GSD_HUE_LT) && (h <= GSD_HUE_UT))
{
if ((h >= nSkinHueLowerBound) && (h <= nSkinHueUpperBound))
ASD_INTENSITY_SET_PIXEL(pFilteredFrame, h);
if (ASD_IS_IN_MOTION(pLastGrayFrame, v, 7))
ASD_INTENSITY_SET_PIXEL(pMotionFrame, h);
}
}
pShrinked += 3;
pGrayFrame++;
pLastGrayFrame++;
pMotionFrame++;
pHueFrame++;
pFilteredFrame++;
}
if (isInit)
cvCalcHist(&imgHueFrame, skinHueHistogram.fHistogram);
cvCopy(imgGrayFrame, imgLastGrayFrame);
cvErode(imgMotionFrame, imgTemp); // eliminate disperse pixels, which occur because of the camera noise
cvDilate(imgTemp, imgMotionFrame);
cvCalcHist(&imgMotionFrame, histogramHueMotion.fHistogram);
skinHueHistogram.mergeWith(&histogramHueMotion, fHistogramMergeFactor);
skinHueHistogram.findCurveThresholds(nSkinHueLowerBound, nSkinHueUpperBound, 1 - fHuePercentCovered);
switch (nMorphingMethod)
{
case MORPHING_METHOD_ERODE :
cvErode(imgFilteredFrame, imgTemp);
cvCopy(imgTemp, imgFilteredFrame);
break;
case MORPHING_METHOD_ERODE_ERODE :
cvErode(imgFilteredFrame, imgTemp);
cvErode(imgTemp, imgFilteredFrame);
break;
case MORPHING_METHOD_ERODE_DILATE :
cvErode(imgFilteredFrame, imgTemp);
cvDilate(imgTemp, imgFilteredFrame);
break;
}
if (outputHueMask != NULL)
cvCopy(imgFilteredFrame, outputHueMask);
}
//------------------------- Histogram for Adaptive Skin Detector -------------------------//
CvAdaptiveSkinDetector::Histogram::Histogram()
{
int histogramSize[] = { HistogramSize };
float range[] = { GSD_HUE_LT, GSD_HUE_UT };
float *ranges[] = { range };
fHistogram = cvCreateHist(1, histogramSize, CV_HIST_ARRAY, ranges, 1);
cvClearHist(fHistogram);
}
CvAdaptiveSkinDetector::Histogram::~Histogram()
{
cvReleaseHist(&fHistogram);
}
int CvAdaptiveSkinDetector::Histogram::findCoverageIndex(double surfaceToCover, int defaultValue)
{
double s = 0;
for (int i = 0; i < HistogramSize; i++)
{
s += cvGetReal1D( fHistogram->bins, i );
if (s >= surfaceToCover)
{
return i;
}
}
return defaultValue;
}
void CvAdaptiveSkinDetector::Histogram::findCurveThresholds(int &x1, int &x2, double percent)
{
double sum = 0;
for (int i = 0; i < HistogramSize; i++)
{
sum += cvGetReal1D( fHistogram->bins, i );
}
x1 = findCoverageIndex(sum * percent, -1);
x2 = findCoverageIndex(sum * (1-percent), -1);
if (x1 == -1)
x1 = GSD_HUE_LT;
else
x1 += GSD_HUE_LT;
if (x2 == -1)
x2 = GSD_HUE_UT;
else
x2 += GSD_HUE_LT;
}
void CvAdaptiveSkinDetector::Histogram::mergeWith(CvAdaptiveSkinDetector::Histogram *source, double weight)
{
float myweight = (float)(1-weight);
float maxVal1 = 0, maxVal2 = 0, *f1, *f2, ff1, ff2;
cvGetMinMaxHistValue(source->fHistogram, NULL, &maxVal2);
if (maxVal2 > 0 )
{
cvGetMinMaxHistValue(fHistogram, NULL, &maxVal1);
if (maxVal1 <= 0)
{
for (int i = 0; i < HistogramSize; i++)
{
f1 = (float*)cvPtr1D(fHistogram->bins, i);
f2 = (float*)cvPtr1D(source->fHistogram->bins, i);
(*f1) = (*f2);
}
}
else
{
for (int i = 0; i < HistogramSize; i++)
{
f1 = (float*)cvPtr1D(fHistogram->bins, i);
f2 = (float*)cvPtr1D(source->fHistogram->bins, i);
ff1 = ((*f1)/maxVal1)*myweight;
if (ff1 < 0)
ff1 = -ff1;
ff2 = (float)(((*f2)/maxVal2)*weight);
if (ff2 < 0)
ff2 = -ff2;
(*f1) = (ff1 + ff2);
}
}
}
}

1128
modules/contrib/src/ba.cpp
View File

File diff suppressed because it is too large Load Diff

138
modules/contrib/src/bowmsctrainer.cpp
View File

@@ -1,138 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
// This file originates from the openFABMAP project:
// [http://code.google.com/p/openfabmap/]
//
// For published work which uses all or part of OpenFABMAP, please cite:
// [http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6224843]
//
// Original Algorithm by Mark Cummins and Paul Newman:
// [http://ijr.sagepub.com/content/27/6/647.short]
// [http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5613942]
// [http://ijr.sagepub.com/content/30/9/1100.abstract]
//
// License Agreement
//
// Copyright (C) 2012 Arren Glover [aj.glover@qut.edu.au] and
// Will Maddern [w.maddern@qut.edu.au], all rights reserved.
//
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "precomp.hpp"
#include "opencv2/contrib/openfabmap.hpp"
namespace cv {
namespace of2 {
BOWMSCTrainer::BOWMSCTrainer(double _clusterSize) :
clusterSize(_clusterSize) {
}
BOWMSCTrainer::~BOWMSCTrainer() {
}
Mat BOWMSCTrainer::cluster() const {
CV_Assert(!descriptors.empty());
int descCount = 0;
for(size_t i = 0; i < descriptors.size(); i++)
descCount += descriptors[i].rows;
Mat mergedDescriptors(descCount, descriptors[0].cols,
descriptors[0].type());
for(size_t i = 0, start = 0; i < descriptors.size(); i++)
{
Mat submut = mergedDescriptors.rowRange((int)start,
(int)(start + descriptors[i].rows));
descriptors[i].copyTo(submut);
start += descriptors[i].rows;
}
return cluster(mergedDescriptors);
}
Mat BOWMSCTrainer::cluster(const Mat& _descriptors) const {
CV_Assert(!_descriptors.empty());
// TODO: sort the descriptors before clustering.
Mat icovar = Mat::eye(_descriptors.cols,_descriptors.cols,_descriptors.type());
std::vector<Mat> initialCentres;
initialCentres.push_back(_descriptors.row(0));
for (int i = 1; i < _descriptors.rows; i++) {
double minDist = DBL_MAX;
for (size_t j = 0; j < initialCentres.size(); j++) {
minDist = std::min(minDist,
cv::Mahalanobis(_descriptors.row(i),initialCentres[j],
icovar));
}
if (minDist > clusterSize)
initialCentres.push_back(_descriptors.row(i));
}
std::vector<std::list<cv::Mat> > clusters;
clusters.resize(initialCentres.size());
for (int i = 0; i < _descriptors.rows; i++) {
int index = 0; double dist = 0, minDist = DBL_MAX;
for (size_t j = 0; j < initialCentres.size(); j++) {
dist = cv::Mahalanobis(_descriptors.row(i),initialCentres[j],icovar);
if (dist < minDist) {
minDist = dist;
index = (int)j;
}
}
clusters[index].push_back(_descriptors.row(i));
}
// TODO: throw away small clusters.
Mat vocabulary;
Mat centre = Mat::zeros(1,_descriptors.cols,_descriptors.type());
for (size_t i = 0; i < clusters.size(); i++) {
centre.setTo(0);
for (std::list<cv::Mat>::iterator Ci = clusters[i].begin(); Ci != clusters[i].end(); Ci++) {
centre += *Ci;
}
centre /= (double)clusters[i].size();
vocabulary.push_back(centre);
}
return vocabulary;
}
}
}

1411
modules/contrib/src/chamfermatching.cpp
View File

File diff suppressed because it is too large Load Diff

289
modules/contrib/src/chowliutree.cpp
View File

@@ -1,289 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
// This file originates from the openFABMAP project:
// [http://code.google.com/p/openfabmap/]
//
// For published work which uses all or part of OpenFABMAP, please cite:
// [http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6224843]
//
// Original Algorithm by Mark Cummins and Paul Newman:
// [http://ijr.sagepub.com/content/27/6/647.short]
// [http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5613942]
// [http://ijr.sagepub.com/content/30/9/1100.abstract]
//
// License Agreement
//
// Copyright (C) 2012 Arren Glover [aj.glover@qut.edu.au] and
// Will Maddern [w.maddern@qut.edu.au], all rights reserved.
//
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "precomp.hpp"
#include "opencv2/contrib/openfabmap.hpp"
namespace cv {
namespace of2 {
ChowLiuTree::ChowLiuTree() {
}
ChowLiuTree::~ChowLiuTree() {
}
void ChowLiuTree::add(const Mat& imgDescriptor) {
CV_Assert(!imgDescriptor.empty());
if (!imgDescriptors.empty()) {
CV_Assert(imgDescriptors[0].cols == imgDescriptor.cols);
CV_Assert(imgDescriptors[0].type() == imgDescriptor.type());
}
imgDescriptors.push_back(imgDescriptor);
}
void ChowLiuTree::add(const std::vector<Mat>& _imgDescriptors) {
for (size_t i = 0; i < _imgDescriptors.size(); i++) {
add(_imgDescriptors[i]);
}
}
const std::vector<cv::Mat>& ChowLiuTree::getImgDescriptors() const {
return imgDescriptors;
}
Mat ChowLiuTree::make(double infoThreshold) {
CV_Assert(!imgDescriptors.empty());
unsigned int descCount = 0;
for (size_t i = 0; i < imgDescriptors.size(); i++)
descCount += imgDescriptors[i].rows;
mergedImgDescriptors = cv::Mat(descCount, imgDescriptors[0].cols,
imgDescriptors[0].type());
for (size_t i = 0, start = 0; i < imgDescriptors.size(); i++)
{
Mat submut = mergedImgDescriptors.rowRange((int)start,
(int)(start + imgDescriptors[i].rows));
imgDescriptors[i].copyTo(submut);
start += imgDescriptors[i].rows;
}
std::list<info> edges;
createBaseEdges(edges, infoThreshold);
// TODO: if it cv_asserts here they really won't know why.
CV_Assert(reduceEdgesToMinSpan(edges));
return buildTree(edges.front().word1, edges);
}
double ChowLiuTree::P(int a, bool za) {
if(za) {
return (0.98 * cv::countNonZero(mergedImgDescriptors.col(a)) /
mergedImgDescriptors.rows) + 0.01;
} else {
return 1 - ((0.98 * cv::countNonZero(mergedImgDescriptors.col(a)) /
mergedImgDescriptors.rows) + 0.01);
}
}
double ChowLiuTree::JP(int a, bool za, int b, bool zb) {
double count = 0;
for(int i = 0; i < mergedImgDescriptors.rows; i++) {
if((mergedImgDescriptors.at<float>(i,a) > 0) == za &&
(mergedImgDescriptors.at<float>(i,b) > 0) == zb) {
count++;
}
}
return count / mergedImgDescriptors.rows;
}
double ChowLiuTree::CP(int a, bool za, int b, bool zb){
int count = 0, total = 0;
for(int i = 0; i < mergedImgDescriptors.rows; i++) {
if((mergedImgDescriptors.at<float>(i,b) > 0) == zb) {
total++;
if((mergedImgDescriptors.at<float>(i,a) > 0) == za) {
count++;
}
}
}
if(total) {
return (double)(0.98 * count)/total + 0.01;
} else {
return (za) ? 0.01 : 0.99;
}
}
cv::Mat ChowLiuTree::buildTree(int root_word, std::list<info> &edges) {
int q = root_word;
cv::Mat cltree(4, (int)edges.size()+1, CV_64F);
cltree.at<double>(0, q) = q;
cltree.at<double>(1, q) = P(q, true);
cltree.at<double>(2, q) = P(q, true);
cltree.at<double>(3, q) = P(q, true);
//setting P(zq|zpq) to P(zq) gives the root node of the chow-liu
//independence from a parent node.
//find all children and do the same
std::vector<int> nextqs = extractChildren(edges, q);
int pq = q;
std::vector<int>::iterator nextq;
for(nextq = nextqs.begin(); nextq != nextqs.end(); nextq++) {
recAddToTree(cltree, *nextq, pq, edges);
}
return cltree;
}
void ChowLiuTree::recAddToTree(cv::Mat &cltree, int q, int pq,
std::list<info>& remaining_edges) {
cltree.at<double>(0, q) = pq;
cltree.at<double>(1, q) = P(q, true);
cltree.at<double>(2, q) = CP(q, true, pq, true);
cltree.at<double>(3, q) = CP(q, true, pq, false);
//find all children and do the same
std::vector<int> nextqs = extractChildren(remaining_edges, q);
pq = q;
std::vector<int>::iterator nextq;
for(nextq = nextqs.begin(); nextq != nextqs.end(); nextq++) {
recAddToTree(cltree, *nextq, pq, remaining_edges);
}
}
std::vector<int> ChowLiuTree::extractChildren(std::list<info> &remaining_edges, int q) {
std::vector<int> children;
std::list<info>::iterator edge = remaining_edges.begin();
while(edge != remaining_edges.end()) {
if(edge->word1 == q) {
children.push_back(edge->word2);
edge = remaining_edges.erase(edge);
continue;
}
if(edge->word2 == q) {
children.push_back(edge->word1);
edge = remaining_edges.erase(edge);
continue;
}
edge++;
}
return children;
}
bool ChowLiuTree::sortInfoScores(const info& first, const info& second) {
return first.score > second.score;
}
double ChowLiuTree::calcMutInfo(int word1, int word2) {
double accumulation = 0;
double P00 = JP(word1, false, word2, false);
if(P00) accumulation += P00 * std::log(P00 / (P(word1, false)*P(word2, false)));
double P01 = JP(word1, false, word2, true);
if(P01) accumulation += P01 * std::log(P01 / (P(word1, false)*P(word2, true)));
double P10 = JP(word1, true, word2, false);
if(P10) accumulation += P10 * std::log(P10 / (P(word1, true)*P(word2, false)));
double P11 = JP(word1, true, word2, true);
if(P11) accumulation += P11 * std::log(P11 / (P(word1, true)*P(word2, true)));
return accumulation;
}
void ChowLiuTree::createBaseEdges(std::list<info>& edges, double infoThreshold) {
int nWords = imgDescriptors[0].cols;
info mutInfo;
for(int word1 = 0; word1 < nWords; word1++) {
for(int word2 = word1 + 1; word2 < nWords; word2++) {
mutInfo.word1 = (short)word1;
mutInfo.word2 = (short)word2;
mutInfo.score = (float)calcMutInfo(word1, word2);
if(mutInfo.score >= infoThreshold)
edges.push_back(mutInfo);
}
}
edges.sort(sortInfoScores);
}
bool ChowLiuTree::reduceEdgesToMinSpan(std::list<info>& edges) {
std::map<int, int> groups;
std::map<int, int>::iterator groupIt;
for(int i = 0; i < imgDescriptors[0].cols; i++) groups[i] = i;
int group1, group2;
std::list<info>::iterator edge = edges.begin();
while(edge != edges.end()) {
if(groups[edge->word1] != groups[edge->word2]) {
group1 = groups[edge->word1];
group2 = groups[edge->word2];
for(groupIt = groups.begin(); groupIt != groups.end(); groupIt++)
if(groupIt->second == group2) groupIt->second = group1;
edge++;
} else {
edge = edges.erase(edge);
}
}
if(edges.size() != (unsigned int)imgDescriptors[0].cols - 1) {
return false;
} else {
return true;
}
}
}
}

530
modules/contrib/src/colormap.cpp
View File

@@ -1,530 +0,0 @@
/*
* Copyright (c) 2011. Philipp Wagner <bytefish[at]gmx[dot]de>.
* Released to public domain under terms of the BSD Simplified license.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of the organization nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* See <http://www.opensource.org/licenses/bsd-license>
*/
#include "precomp.hpp"
#include <iostream>
#ifdef _MSC_VER
#pragma warning( disable: 4305 )
#endif
namespace cv
{
static Mat linspace(float x0, float x1, int n)
{
Mat pts(n, 1, CV_32FC1);
float step = (x1-x0)/(n-1);
for(int i = 0; i < n; i++)
pts.at<float>(i,0) = x0+i*step;
return pts;
}
//------------------------------------------------------------------------------
// cv::sortMatrixRowsByIndices
//------------------------------------------------------------------------------
static void sortMatrixRowsByIndices(InputArray _src, InputArray _indices, OutputArray _dst)
{
if(_indices.getMat().type() != CV_32SC1)
CV_Error(Error::StsUnsupportedFormat, "cv::sortRowsByIndices only works on integer indices!");
Mat src = _src.getMat();
std::vector<int> indices = _indices.getMat();
_dst.create(src.rows, src.cols, src.type());
Mat dst = _dst.getMat();
for(size_t idx = 0; idx < indices.size(); idx++) {
Mat originalRow = src.row(indices[idx]);
Mat sortedRow = dst.row((int)idx);
originalRow.copyTo(sortedRow);
}
}
static Mat sortMatrixRowsByIndices(InputArray src, InputArray indices)
{
Mat dst;
sortMatrixRowsByIndices(src, indices, dst);
return dst;
}
static Mat argsort(InputArray _src, bool ascending=true)
{
Mat src = _src.getMat();
if (src.rows != 1 && src.cols != 1)
CV_Error(Error::StsBadArg, "cv::argsort only sorts 1D matrices.");
int flags = SORT_EVERY_ROW | (ascending ? SORT_ASCENDING : SORT_DESCENDING);
Mat sorted_indices;
sortIdx(src.reshape(1,1),sorted_indices,flags);
return sorted_indices;
}
template <typename _Tp> static
Mat interp1_(const Mat& X_, const Mat& Y_, const Mat& XI)
{
int n = XI.rows;
// sort input table
std::vector<int> sort_indices = argsort(X_);
Mat X = sortMatrixRowsByIndices(X_,sort_indices);
Mat Y = sortMatrixRowsByIndices(Y_,sort_indices);
// interpolated values
Mat yi = Mat::zeros(XI.size(), XI.type());
for(int i = 0; i < n; i++) {
int c = 0;
int low = 0;
int high = X.rows - 1;
// set bounds
if(XI.at<_Tp>(i,0) < X.at<_Tp>(low, 0))
high = 1;
if(XI.at<_Tp>(i,0) > X.at<_Tp>(high, 0))
low = high - 1;
// binary search
while((high-low)>1) {
c = low + ((high - low) >> 1);
if(XI.at<_Tp>(i,0) > X.at<_Tp>(c,0)) {
low = c;
} else {
high = c;
}
}
// linear interpolation
yi.at<_Tp>(i,0) += Y.at<_Tp>(low,0)
+ (XI.at<_Tp>(i,0) - X.at<_Tp>(low,0))
* (Y.at<_Tp>(high,0) - Y.at<_Tp>(low,0))
/ (X.at<_Tp>(high,0) - X.at<_Tp>(low,0));
}
return yi;
}
static Mat interp1(InputArray _x, InputArray _Y, InputArray _xi)
{
// get matrices
Mat x = _x.getMat();
Mat Y = _Y.getMat();
Mat xi = _xi.getMat();
// check types & alignment
CV_Assert((x.type() == Y.type()) && (Y.type() == xi.type()));
CV_Assert((x.cols == 1) && (x.rows == Y.rows) && (x.cols == Y.cols));
// call templated interp1
switch(x.type()) {
case CV_8SC1: return interp1_<char>(x,Y,xi); break;
case CV_8UC1: return interp1_<unsigned char>(x,Y,xi); break;
case CV_16SC1: return interp1_<short>(x,Y,xi); break;
case CV_16UC1: return interp1_<unsigned short>(x,Y,xi); break;
case CV_32SC1: return interp1_<int>(x,Y,xi); break;
case CV_32FC1: return interp1_<float>(x,Y,xi); break;
case CV_64FC1: return interp1_<double>(x,Y,xi); break;
default: CV_Error(Error::StsUnsupportedFormat, ""); break;
}
return Mat();
}
namespace colormap
{
class ColorMap {
protected:
Mat _lut;
public:
virtual ~ColorMap() {}
// Applies the colormap on a given image.
//
// This function expects BGR-aligned data of type CV_8UC1 or
// CV_8UC3. If the wrong image type is given, the original image
// will be returned.
//
// Throws an error for wrong-aligned lookup table, which must be
// of size 256 in the latest OpenCV release (2.3.1).
void operator()(InputArray src, OutputArray dst) const;
// Setup base map to interpolate from.
virtual void init(int n) = 0;
// Interpolates from a base colormap.
static Mat linear_colormap(InputArray X,
InputArray r, InputArray g, InputArray b,
int n) {
return linear_colormap(X,r,g,b,linspace(0,1,n));
}
// Interpolates from a base colormap.
static Mat linear_colormap(InputArray X,
InputArray r, InputArray g, InputArray b,
float begin, float end, float n) {
return linear_colormap(X,r,g,b,linspace(begin,end, cvRound(n)));
}
// Interpolates from a base colormap.
static Mat linear_colormap(InputArray X,
InputArray r, InputArray g, InputArray b,
InputArray xi);
};
// Equals the GNU Octave colormap "autumn".
class Autumn : public ColorMap {
public:
Autumn() : ColorMap() {
init(256);
}
Autumn(int n) : ColorMap() {
init(n);
}
void init(int n) {
float r[] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
float g[] = { 0, 0.01587301587301587, 0.03174603174603174, 0.04761904761904762, 0.06349206349206349, 0.07936507936507936, 0.09523809523809523, 0.1111111111111111, 0.126984126984127, 0.1428571428571428, 0.1587301587301587, 0.1746031746031746, 0.1904761904761905, 0.2063492063492063, 0.2222222222222222, 0.2380952380952381, 0.253968253968254, 0.2698412698412698, 0.2857142857142857, 0.3015873015873016, 0.3174603174603174, 0.3333333333333333, 0.3492063492063492, 0.3650793650793651, 0.3809523809523809, 0.3968253968253968, 0.4126984126984127, 0.4285714285714285, 0.4444444444444444, 0.4603174603174603, 0.4761904761904762, 0.492063492063492, 0.5079365079365079, 0.5238095238095238, 0.5396825396825397, 0.5555555555555556, 0.5714285714285714, 0.5873015873015873, 0.6031746031746031, 0.6190476190476191, 0.6349206349206349, 0.6507936507936508, 0.6666666666666666, 0.6825396825396826, 0.6984126984126984, 0.7142857142857143, 0.7301587301587301, 0.746031746031746, 0.7619047619047619, 0.7777777777777778, 0.7936507936507936, 0.8095238095238095, 0.8253968253968254, 0.8412698412698413, 0.8571428571428571, 0.873015873015873, 0.8888888888888888, 0.9047619047619048, 0.9206349206349206, 0.9365079365079365, 0.9523809523809523, 0.9682539682539683, 0.9841269841269841, 1};
float b[] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
Mat X = linspace(0,1,64);
this->_lut = ColorMap::linear_colormap(X,
Mat(64,1, CV_32FC1, r).clone(), // red
Mat(64,1, CV_32FC1, g).clone(), // green
Mat(64,1, CV_32FC1, b).clone(), // blue
n); // number of sample points
}
};
// Equals the GNU Octave colormap "bone".
class Bone : public ColorMap {
public:
Bone() : ColorMap() {
init(256);
}
Bone(int n) : ColorMap() {
init(n);
}
void init(int n) {
float r[] = { 0, 0.01388888888888889, 0.02777777777777778, 0.04166666666666666, 0.05555555555555555, 0.06944444444444445, 0.08333333333333333, 0.09722222222222221, 0.1111111111111111, 0.125, 0.1388888888888889, 0.1527777777777778, 0.1666666666666667, 0.1805555555555556, 0.1944444444444444, 0.2083333333333333, 0.2222222222222222, 0.2361111111111111, 0.25, 0.2638888888888889, 0.2777777777777778, 0.2916666666666666, 0.3055555555555555, 0.3194444444444444, 0.3333333333333333, 0.3472222222222222, 0.3611111111111111, 0.375, 0.3888888888888888, 0.4027777777777777, 0.4166666666666666, 0.4305555555555555, 0.4444444444444444, 0.4583333333333333, 0.4722222222222222, 0.4861111111111112, 0.5, 0.5138888888888888, 0.5277777777777778, 0.5416666666666667, 0.5555555555555556, 0.5694444444444444, 0.5833333333333333, 0.5972222222222222, 0.611111111111111, 0.6249999999999999, 0.6388888888888888, 0.6527777777777778, 0.6726190476190474, 0.6944444444444442, 0.7162698412698412, 0.7380952380952381, 0.7599206349206349, 0.7817460317460316, 0.8035714285714286, 0.8253968253968254, 0.8472222222222221, 0.8690476190476188, 0.8908730158730158, 0.9126984126984128, 0.9345238095238095, 0.9563492063492063, 0.978174603174603, 1};
float g[] = { 0, 0.01388888888888889, 0.02777777777777778, 0.04166666666666666, 0.05555555555555555, 0.06944444444444445, 0.08333333333333333, 0.09722222222222221, 0.1111111111111111, 0.125, 0.1388888888888889, 0.1527777777777778, 0.1666666666666667, 0.1805555555555556, 0.1944444444444444, 0.2083333333333333, 0.2222222222222222, 0.2361111111111111, 0.25, 0.2638888888888889, 0.2777777777777778, 0.2916666666666666, 0.3055555555555555, 0.3194444444444444, 0.3353174603174602, 0.3544973544973544, 0.3736772486772486, 0.3928571428571428, 0.412037037037037, 0.4312169312169312, 0.4503968253968254, 0.4695767195767195, 0.4887566137566137, 0.5079365079365078, 0.5271164021164021, 0.5462962962962963, 0.5654761904761904, 0.5846560846560845, 0.6038359788359787, 0.623015873015873, 0.6421957671957671, 0.6613756613756612, 0.6805555555555555, 0.6997354497354497, 0.7189153439153438, 0.7380952380952379, 0.7572751322751322, 0.7764550264550264, 0.7916666666666666, 0.8055555555555555, 0.8194444444444444, 0.8333333333333334, 0.8472222222222222, 0.861111111111111, 0.875, 0.8888888888888888, 0.9027777777777777, 0.9166666666666665, 0.9305555555555555, 0.9444444444444444, 0.9583333333333333, 0.9722222222222221, 0.986111111111111, 1};
float b[] = { 0, 0.01917989417989418, 0.03835978835978836, 0.05753968253968253, 0.07671957671957672, 0.09589947089947089, 0.1150793650793651, 0.1342592592592592, 0.1534391534391534, 0.1726190476190476, 0.1917989417989418, 0.210978835978836, 0.2301587301587301, 0.2493386243386243, 0.2685185185185185, 0.2876984126984127, 0.3068783068783069, 0.326058201058201, 0.3452380952380952, 0.3644179894179894, 0.3835978835978835, 0.4027777777777777, 0.4219576719576719, 0.4411375661375661, 0.4583333333333333, 0.4722222222222222, 0.4861111111111111, 0.5, 0.5138888888888888, 0.5277777777777777, 0.5416666666666666, 0.5555555555555556, 0.5694444444444444, 0.5833333333333333, 0.5972222222222222, 0.6111111111111112, 0.625, 0.6388888888888888, 0.6527777777777778, 0.6666666666666667, 0.6805555555555556, 0.6944444444444444, 0.7083333333333333, 0.7222222222222222, 0.736111111111111, 0.7499999999999999, 0.7638888888888888, 0.7777777777777778, 0.7916666666666666, 0.8055555555555555, 0.8194444444444444, 0.8333333333333334, 0.8472222222222222, 0.861111111111111, 0.875, 0.8888888888888888, 0.9027777777777777, 0.9166666666666665, 0.9305555555555555, 0.9444444444444444, 0.9583333333333333, 0.9722222222222221, 0.986111111111111, 1};
Mat X = linspace(0,1,64);
this->_lut = ColorMap::linear_colormap(X,
Mat(64,1, CV_32FC1, r).clone(), // red
Mat(64,1, CV_32FC1, g).clone(), // green
Mat(64,1, CV_32FC1, b).clone(), // blue
n); // number of sample points
}
};
// Equals the GNU Octave colormap "jet".
class Jet : public ColorMap {
public:
Jet() {
init(256);
}
Jet(int n) : ColorMap() {
init(n);
}
void init(int n) {
// breakpoints
Mat X = linspace(0,1,256);
// define the basemap
float r[] = {0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.00588235294117645,0.02156862745098032,0.03725490196078418,0.05294117647058827,0.06862745098039214,0.084313725490196,0.1000000000000001,0.115686274509804,0.1313725490196078,0.1470588235294117,0.1627450980392156,0.1784313725490196,0.1941176470588235,0.2098039215686274,0.2254901960784315,0.2411764705882353,0.2568627450980392,0.2725490196078431,0.2882352941176469,0.303921568627451,0.3196078431372549,0.3352941176470587,0.3509803921568628,0.3666666666666667,0.3823529411764706,0.3980392156862744,0.4137254901960783,0.4294117647058824,0.4450980392156862,0.4607843137254901,0.4764705882352942,0.4921568627450981,0.5078431372549019,0.5235294117647058,0.5392156862745097,0.5549019607843135,0.5705882352941174,0.5862745098039217,0.6019607843137256,0.6176470588235294,0.6333333333333333,0.6490196078431372,0.664705882352941,0.6803921568627449,0.6960784313725492,0.7117647058823531,0.7274509803921569,0.7431372549019608,0.7588235294117647,0.7745098039215685,0.7901960784313724,0.8058823529411763,0.8215686274509801,0.8372549019607844,0.8529411764705883,0.8686274509803922,0.884313725490196,0.8999999999999999,0.9156862745098038,0.9313725490196076,0.947058823529412,0.9627450980392158,0.9784313725490197,0.9941176470588236,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0.9862745098039216,0.9705882352941178,0.9549019607843139,0.93921568627451,0.9235294117647062,0.9078431372549018,0.892156862745098,0.8764705882352941,0.8607843137254902,0.8450980392156864,0.8294117647058825,0.8137254901960786,0.7980392156862743,0.7823529411764705,0.7666666666666666,0.7509803921568627,0.7352941176470589,0.719607843137255,0.7039215686274511,0.6882352941176473,0.6725490196078434,0.6568627450980391,0.6411764705882352,0.6254901960784314,0.6098039215686275,0.5941176470588236,0.5784313725490198,0.5627450980392159,0.5470588235294116,0.5313725490196077,0.5156862745098039,0.5};
float g[] = {0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.001960784313725483,0.01764705882352935,0.03333333333333333,0.0490196078431373,0.06470588235294117,0.08039215686274503,0.09607843137254901,0.111764705882353,0.1274509803921569,0.1431372549019607,0.1588235294117647,0.1745098039215687,0.1901960784313725,0.2058823529411764,0.2215686274509804,0.2372549019607844,0.2529411764705882,0.2686274509803921,0.2843137254901961,0.3,0.3156862745098039,0.3313725490196078,0.3470588235294118,0.3627450980392157,0.3784313725490196,0.3941176470588235,0.4098039215686274,0.4254901960784314,0.4411764705882353,0.4568627450980391,0.4725490196078431,0.4882352941176471,0.503921568627451,0.5196078431372548,0.5352941176470587,0.5509803921568628,0.5666666666666667,0.5823529411764705,0.5980392156862746,0.6137254901960785,0.6294117647058823,0.6450980392156862,0.6607843137254901,0.6764705882352942,0.692156862745098,0.7078431372549019,0.723529411764706,0.7392156862745098,0.7549019607843137,0.7705882352941176,0.7862745098039214,0.8019607843137255,0.8176470588235294,0.8333333333333333,0.8490196078431373,0.8647058823529412,0.8803921568627451,0.8960784313725489,0.9117647058823528,0.9274509803921569,0.9431372549019608,0.9588235294117646,0.9745098039215687,0.9901960784313726,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0.9901960784313726,0.9745098039215687,0.9588235294117649,0.943137254901961,0.9274509803921571,0.9117647058823528,0.8960784313725489,0.8803921568627451,0.8647058823529412,0.8490196078431373,0.8333333333333335,0.8176470588235296,0.8019607843137253,0.7862745098039214,0.7705882352941176,0.7549019607843137,0.7392156862745098,0.723529411764706,0.7078431372549021,0.6921568627450982,0.6764705882352944,0.6607843137254901,0.6450980392156862,0.6294117647058823,0.6137254901960785,0.5980392156862746,0.5823529411764707,0.5666666666666669,0.5509803921568626,0.5352941176470587,0.5196078431372548,0.503921568627451,0.4882352941176471,0.4725490196078432,0.4568627450980394,0.4411764705882355,0.4254901960784316,0.4098039215686273,0.3941176470588235,0.3784313725490196,0.3627450980392157,0.3470588235294119,0.331372549019608,0.3156862745098041,0.2999999999999998,0.284313725490196,0.2686274509803921,0.2529411764705882,0.2372549019607844,0.2215686274509805,0.2058823529411766,0.1901960784313728,0.1745098039215689,0.1588235294117646,0.1431372549019607,0.1274509803921569,0.111764705882353,0.09607843137254912,0.08039215686274526,0.06470588235294139,0.04901960784313708,0.03333333333333321,0.01764705882352935,0.001960784313725483,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0};
float b[] = {0.5,0.5156862745098039,0.5313725490196078,0.5470588235294118,0.5627450980392157,0.5784313725490196,0.5941176470588235,0.6098039215686275,0.6254901960784314,0.6411764705882352,0.6568627450980392,0.6725490196078432,0.6882352941176471,0.7039215686274509,0.7196078431372549,0.7352941176470589,0.7509803921568627,0.7666666666666666,0.7823529411764706,0.7980392156862746,0.8137254901960784,0.8294117647058823,0.8450980392156863,0.8607843137254902,0.8764705882352941,0.892156862745098,0.907843137254902,0.9235294117647059,0.9392156862745098,0.9549019607843137,0.9705882352941176,0.9862745098039216,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0.9941176470588236,0.9784313725490197,0.9627450980392158,0.9470588235294117,0.9313725490196079,0.915686274509804,0.8999999999999999,0.884313725490196,0.8686274509803922,0.8529411764705883,0.8372549019607844,0.8215686274509804,0.8058823529411765,0.7901960784313726,0.7745098039215685,0.7588235294117647,0.7431372549019608,0.7274509803921569,0.7117647058823531,0.696078431372549,0.6803921568627451,0.6647058823529413,0.6490196078431372,0.6333333333333333,0.6176470588235294,0.6019607843137256,0.5862745098039217,0.5705882352941176,0.5549019607843138,0.5392156862745099,0.5235294117647058,0.5078431372549019,0.4921568627450981,0.4764705882352942,0.4607843137254903,0.4450980392156865,0.4294117647058826,0.4137254901960783,0.3980392156862744,0.3823529411764706,0.3666666666666667,0.3509803921568628,0.335294117647059,0.3196078431372551,0.3039215686274508,0.2882352941176469,0.2725490196078431,0.2568627450980392,0.2411764705882353,0.2254901960784315,0.2098039215686276,0.1941176470588237,0.1784313725490199,0.1627450980392156,0.1470588235294117,0.1313725490196078,0.115686274509804,0.1000000000000001,0.08431372549019622,0.06862745098039236,0.05294117647058805,0.03725490196078418,0.02156862745098032,0.00588235294117645,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0};
// now build lookup table
this->_lut = ColorMap::linear_colormap(X,
Mat(256,1, CV_32FC1, r).clone(), // red
Mat(256,1, CV_32FC1, g).clone(), // green
Mat(256,1, CV_32FC1, b).clone(), // blue
n);
}
};
// Equals the GNU Octave colormap "winter".
class Winter : public ColorMap {
public:
Winter() : ColorMap() {
init(256);
}
Winter(int n) : ColorMap() {
init(n);
}
void init(int n) {
float r[] = {0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0};
float g[] = {0.0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0};
float b[] = {1.0, 0.95, 0.9, 0.85, 0.8, 0.75, 0.7, 0.65, 0.6, 0.55, 0.5};
Mat X = linspace(0,1,11);
this->_lut = ColorMap::linear_colormap(X,
Mat(11,1, CV_32FC1, r).clone(), // red
Mat(11,1, CV_32FC1, g).clone(), // green
Mat(11,1, CV_32FC1, b).clone(), // blue
n); // number of sample points
}
};
// Equals the GNU Octave colormap "rainbow".
class Rainbow : public ColorMap {
public:
Rainbow() : ColorMap() {
init(256);
}
Rainbow(int n) : ColorMap() {
init(n);
}
void init(int n) {
float r[] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0.9365079365079367, 0.8571428571428572, 0.7777777777777777, 0.6984126984126986, 0.6190476190476191, 0.53968253968254, 0.4603174603174605, 0.3809523809523814, 0.3015873015873018, 0.2222222222222223, 0.1428571428571432, 0.06349206349206415, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.03174603174603208, 0.08465608465608465, 0.1375661375661377, 0.1904761904761907, 0.2433862433862437, 0.2962962962962963, 0.3492063492063493, 0.4021164021164023, 0.4550264550264553, 0.5079365079365079, 0.5608465608465609, 0.6137566137566139, 0.666666666666667};
float g[] = { 0, 0.03968253968253968, 0.07936507936507936, 0.119047619047619, 0.1587301587301587, 0.1984126984126984, 0.2380952380952381, 0.2777777777777778, 0.3174603174603174, 0.3571428571428571, 0.3968253968253968, 0.4365079365079365, 0.4761904761904762, 0.5158730158730158, 0.5555555555555556, 0.5952380952380952, 0.6349206349206349, 0.6746031746031745, 0.7142857142857142, 0.753968253968254, 0.7936507936507936, 0.8333333333333333, 0.873015873015873, 0.9126984126984127, 0.9523809523809523, 0.992063492063492, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0.9841269841269842, 0.9047619047619047, 0.8253968253968256, 0.7460317460317465, 0.666666666666667, 0.587301587301587, 0.5079365079365079, 0.4285714285714288, 0.3492063492063493, 0.2698412698412698, 0.1904761904761907, 0.1111111111111116, 0.03174603174603208, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
float b[] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.01587301587301582, 0.09523809523809534, 0.1746031746031744, 0.2539682539682535, 0.333333333333333, 0.412698412698413, 0.4920634920634921, 0.5714285714285712, 0.6507936507936507, 0.7301587301587302, 0.8095238095238093, 0.8888888888888884, 0.9682539682539679, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
Mat X = linspace(0,1,64);
this->_lut = ColorMap::linear_colormap(X,
Mat(64,1, CV_32FC1, r).clone(), // red
Mat(64,1, CV_32FC1, g).clone(), // green
Mat(64,1, CV_32FC1, b).clone(), // blue
n); // number of sample points
}
};
// Equals the GNU Octave colormap "ocean".
class Ocean : public ColorMap {
public:
Ocean() : ColorMap() {
init(256);
}
Ocean(int n) : ColorMap() {
init(n);
}
void init(int n) {
float r[] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.04761904761904762, 0.09523809523809523, 0.1428571428571428, 0.1904761904761905, 0.2380952380952381, 0.2857142857142857, 0.3333333333333333, 0.3809523809523809, 0.4285714285714285, 0.4761904761904762, 0.5238095238095238, 0.5714285714285714, 0.6190476190476191, 0.6666666666666666, 0.7142857142857143, 0.7619047619047619, 0.8095238095238095, 0.8571428571428571, 0.9047619047619048, 0.9523809523809523, 1};
float g[] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.02380952380952381, 0.04761904761904762, 0.07142857142857142, 0.09523809523809523, 0.119047619047619, 0.1428571428571428, 0.1666666666666667, 0.1904761904761905, 0.2142857142857143, 0.2380952380952381, 0.2619047619047619, 0.2857142857142857, 0.3095238095238095, 0.3333333333333333, 0.3571428571428572, 0.3809523809523809, 0.4047619047619048, 0.4285714285714285, 0.4523809523809524, 0.4761904761904762, 0.5, 0.5238095238095238, 0.5476190476190477, 0.5714285714285714, 0.5952380952380952, 0.6190476190476191, 0.6428571428571429, 0.6666666666666666, 0.6904761904761905, 0.7142857142857143, 0.7380952380952381, 0.7619047619047619, 0.7857142857142857, 0.8095238095238095, 0.8333333333333334, 0.8571428571428571, 0.8809523809523809, 0.9047619047619048, 0.9285714285714286, 0.9523809523809523, 0.9761904761904762, 1};
float b[] = { 0, 0.01587301587301587, 0.03174603174603174, 0.04761904761904762, 0.06349206349206349, 0.07936507936507936, 0.09523809523809523, 0.1111111111111111, 0.126984126984127, 0.1428571428571428, 0.1587301587301587, 0.1746031746031746, 0.1904761904761905, 0.2063492063492063, 0.2222222222222222, 0.2380952380952381, 0.253968253968254, 0.2698412698412698, 0.2857142857142857, 0.3015873015873016, 0.3174603174603174, 0.3333333333333333, 0.3492063492063492, 0.3650793650793651, 0.3809523809523809, 0.3968253968253968, 0.4126984126984127, 0.4285714285714285, 0.4444444444444444, 0.4603174603174603, 0.4761904761904762, 0.492063492063492, 0.5079365079365079, 0.5238095238095238, 0.5396825396825397, 0.5555555555555556, 0.5714285714285714, 0.5873015873015873, 0.6031746031746031, 0.6190476190476191, 0.6349206349206349, 0.6507936507936508, 0.6666666666666666, 0.6825396825396826, 0.6984126984126984, 0.7142857142857143, 0.7301587301587301, 0.746031746031746, 0.7619047619047619, 0.7777777777777778, 0.7936507936507936, 0.8095238095238095, 0.8253968253968254, 0.8412698412698413, 0.8571428571428571, 0.873015873015873, 0.8888888888888888, 0.9047619047619048, 0.9206349206349206, 0.9365079365079365, 0.9523809523809523, 0.9682539682539683, 0.9841269841269841, 1};
Mat X = linspace(0,1,64);
this->_lut = ColorMap::linear_colormap(X,
Mat(64,1, CV_32FC1, r).clone(), // red
Mat(64,1, CV_32FC1, g).clone(), // green
Mat(64,1, CV_32FC1, b).clone(), // blue
n); // number of sample points
}
};
// Equals the GNU Octave colormap "summer".
class Summer : public ColorMap {
public:
Summer() : ColorMap() {
init(256);
}
Summer(int n) : ColorMap() {
init(n);
}
void init(int n) {
float r[] = { 0, 0.01587301587301587, 0.03174603174603174, 0.04761904761904762, 0.06349206349206349, 0.07936507936507936, 0.09523809523809523, 0.1111111111111111, 0.126984126984127, 0.1428571428571428, 0.1587301587301587, 0.1746031746031746, 0.1904761904761905, 0.2063492063492063, 0.2222222222222222, 0.2380952380952381, 0.253968253968254, 0.2698412698412698, 0.2857142857142857, 0.3015873015873016, 0.3174603174603174, 0.3333333333333333, 0.3492063492063492, 0.3650793650793651, 0.3809523809523809, 0.3968253968253968, 0.4126984126984127, 0.4285714285714285, 0.4444444444444444, 0.4603174603174603, 0.4761904761904762, 0.492063492063492, 0.5079365079365079, 0.5238095238095238, 0.5396825396825397, 0.5555555555555556, 0.5714285714285714, 0.5873015873015873, 0.6031746031746031, 0.6190476190476191, 0.6349206349206349, 0.6507936507936508, 0.6666666666666666, 0.6825396825396826, 0.6984126984126984, 0.7142857142857143, 0.7301587301587301, 0.746031746031746, 0.7619047619047619, 0.7777777777777778, 0.7936507936507936, 0.8095238095238095, 0.8253968253968254, 0.8412698412698413, 0.8571428571428571, 0.873015873015873, 0.8888888888888888, 0.9047619047619048, 0.9206349206349206, 0.9365079365079365, 0.9523809523809523, 0.9682539682539683, 0.9841269841269841, 1};
float g[] = { 0.5, 0.5079365079365079, 0.5158730158730158, 0.5238095238095238, 0.5317460317460317, 0.5396825396825397, 0.5476190476190477, 0.5555555555555556, 0.5634920634920635, 0.5714285714285714, 0.5793650793650793, 0.5873015873015873, 0.5952380952380952, 0.6031746031746031, 0.6111111111111112, 0.6190476190476191, 0.626984126984127, 0.6349206349206349, 0.6428571428571428, 0.6507936507936508, 0.6587301587301587, 0.6666666666666666, 0.6746031746031746, 0.6825396825396826, 0.6904761904761905, 0.6984126984126984, 0.7063492063492063, 0.7142857142857143, 0.7222222222222222, 0.7301587301587301, 0.7380952380952381, 0.746031746031746, 0.753968253968254, 0.7619047619047619, 0.7698412698412698, 0.7777777777777778, 0.7857142857142857, 0.7936507936507937, 0.8015873015873016, 0.8095238095238095, 0.8174603174603174, 0.8253968253968254, 0.8333333333333333, 0.8412698412698413, 0.8492063492063492, 0.8571428571428572, 0.8650793650793651, 0.873015873015873, 0.8809523809523809, 0.8888888888888888, 0.8968253968253967, 0.9047619047619048, 0.9126984126984127, 0.9206349206349207, 0.9285714285714286, 0.9365079365079365, 0.9444444444444444, 0.9523809523809523, 0.9603174603174602, 0.9682539682539683, 0.9761904761904762, 0.9841269841269842, 0.9920634920634921, 1};
float b[] = { 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4, 0.4};
Mat X = linspace(0,1,64);
this->_lut = ColorMap::linear_colormap(X,
Mat(64,1, CV_32FC1, r).clone(), // red
Mat(64,1, CV_32FC1, g).clone(), // green
Mat(64,1, CV_32FC1, b).clone(), // blue
n); // number of sample points
}
};
// Equals the GNU Octave colormap "spring".
class Spring : public ColorMap {
public:
Spring() : ColorMap() {
init(256);
}
Spring(int n) : ColorMap() {
init(n);
}
void init(int n) {
float r[] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
float g[] = { 0, 0.01587301587301587, 0.03174603174603174, 0.04761904761904762, 0.06349206349206349, 0.07936507936507936, 0.09523809523809523, 0.1111111111111111, 0.126984126984127, 0.1428571428571428, 0.1587301587301587, 0.1746031746031746, 0.1904761904761905, 0.2063492063492063, 0.2222222222222222, 0.2380952380952381, 0.253968253968254, 0.2698412698412698, 0.2857142857142857, 0.3015873015873016, 0.3174603174603174, 0.3333333333333333, 0.3492063492063492, 0.3650793650793651, 0.3809523809523809, 0.3968253968253968, 0.4126984126984127, 0.4285714285714285, 0.4444444444444444, 0.4603174603174603, 0.4761904761904762, 0.492063492063492, 0.5079365079365079, 0.5238095238095238, 0.5396825396825397, 0.5555555555555556, 0.5714285714285714, 0.5873015873015873, 0.6031746031746031, 0.6190476190476191, 0.6349206349206349, 0.6507936507936508, 0.6666666666666666, 0.6825396825396826, 0.6984126984126984, 0.7142857142857143, 0.7301587301587301, 0.746031746031746, 0.7619047619047619, 0.7777777777777778, 0.7936507936507936, 0.8095238095238095, 0.8253968253968254, 0.8412698412698413, 0.8571428571428571, 0.873015873015873, 0.8888888888888888, 0.9047619047619048, 0.9206349206349206, 0.9365079365079365, 0.9523809523809523, 0.9682539682539683, 0.9841269841269841, 1};
float b[] = { 1, 0.9841269841269842, 0.9682539682539683, 0.9523809523809523, 0.9365079365079365, 0.9206349206349207, 0.9047619047619048, 0.8888888888888888, 0.873015873015873, 0.8571428571428572, 0.8412698412698413, 0.8253968253968254, 0.8095238095238095, 0.7936507936507937, 0.7777777777777778, 0.7619047619047619, 0.746031746031746, 0.7301587301587302, 0.7142857142857143, 0.6984126984126984, 0.6825396825396826, 0.6666666666666667, 0.6507936507936508, 0.6349206349206349, 0.6190476190476191, 0.6031746031746033, 0.5873015873015873, 0.5714285714285714, 0.5555555555555556, 0.5396825396825398, 0.5238095238095238, 0.5079365079365079, 0.4920634920634921, 0.4761904761904762, 0.4603174603174603, 0.4444444444444444, 0.4285714285714286, 0.4126984126984127, 0.3968253968253969, 0.3809523809523809, 0.3650793650793651, 0.3492063492063492, 0.3333333333333334, 0.3174603174603174, 0.3015873015873016, 0.2857142857142857, 0.2698412698412699, 0.253968253968254, 0.2380952380952381, 0.2222222222222222, 0.2063492063492064, 0.1904761904761905, 0.1746031746031746, 0.1587301587301587, 0.1428571428571429, 0.126984126984127, 0.1111111111111112, 0.09523809523809523, 0.07936507936507942, 0.06349206349206349, 0.04761904761904767, 0.03174603174603174, 0.01587301587301593, 0};
Mat X = linspace(0,1,64);
this->_lut = ColorMap::linear_colormap(X,
Mat(64,1, CV_32FC1, r).clone(), // red
Mat(64,1, CV_32FC1, g).clone(), // green
Mat(64,1, CV_32FC1, b).clone(), // blue
n); // number of sample points
}
};
// Equals the GNU Octave colormap "cool".
class Cool : public ColorMap {
public:
Cool() : ColorMap() {
init(256);
}
Cool(int n) : ColorMap() {
init(n);
}
void init(int n) {
float r[] = { 0, 0.01587301587301587, 0.03174603174603174, 0.04761904761904762, 0.06349206349206349, 0.07936507936507936, 0.09523809523809523, 0.1111111111111111, 0.126984126984127, 0.1428571428571428, 0.1587301587301587, 0.1746031746031746, 0.1904761904761905, 0.2063492063492063, 0.2222222222222222, 0.2380952380952381, 0.253968253968254, 0.2698412698412698, 0.2857142857142857, 0.3015873015873016, 0.3174603174603174, 0.3333333333333333, 0.3492063492063492, 0.3650793650793651, 0.3809523809523809, 0.3968253968253968, 0.4126984126984127, 0.4285714285714285, 0.4444444444444444, 0.4603174603174603, 0.4761904761904762, 0.492063492063492, 0.5079365079365079, 0.5238095238095238, 0.5396825396825397, 0.5555555555555556, 0.5714285714285714, 0.5873015873015873, 0.6031746031746031, 0.6190476190476191, 0.6349206349206349, 0.6507936507936508, 0.6666666666666666, 0.6825396825396826, 0.6984126984126984, 0.7142857142857143, 0.7301587301587301, 0.746031746031746, 0.7619047619047619, 0.7777777777777778, 0.7936507936507936, 0.8095238095238095, 0.8253968253968254, 0.8412698412698413, 0.8571428571428571, 0.873015873015873, 0.8888888888888888, 0.9047619047619048, 0.9206349206349206, 0.9365079365079365, 0.9523809523809523, 0.9682539682539683, 0.9841269841269841, 1};
float g[] = { 1, 0.9841269841269842, 0.9682539682539683, 0.9523809523809523, 0.9365079365079365, 0.9206349206349207, 0.9047619047619048, 0.8888888888888888, 0.873015873015873, 0.8571428571428572, 0.8412698412698413, 0.8253968253968254, 0.8095238095238095, 0.7936507936507937, 0.7777777777777778, 0.7619047619047619, 0.746031746031746, 0.7301587301587302, 0.7142857142857143, 0.6984126984126984, 0.6825396825396826, 0.6666666666666667, 0.6507936507936508, 0.6349206349206349, 0.6190476190476191, 0.6031746031746033, 0.5873015873015873, 0.5714285714285714, 0.5555555555555556, 0.5396825396825398, 0.5238095238095238, 0.5079365079365079, 0.4920634920634921, 0.4761904761904762, 0.4603174603174603, 0.4444444444444444, 0.4285714285714286, 0.4126984126984127, 0.3968253968253969, 0.3809523809523809, 0.3650793650793651, 0.3492063492063492, 0.3333333333333334, 0.3174603174603174, 0.3015873015873016, 0.2857142857142857, 0.2698412698412699, 0.253968253968254, 0.2380952380952381, 0.2222222222222222, 0.2063492063492064, 0.1904761904761905, 0.1746031746031746, 0.1587301587301587, 0.1428571428571429, 0.126984126984127, 0.1111111111111112, 0.09523809523809523, 0.07936507936507942, 0.06349206349206349, 0.04761904761904767, 0.03174603174603174, 0.01587301587301593, 0};
float b[] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
Mat X = linspace(0,1,64);
this->_lut = ColorMap::linear_colormap(X,
Mat(64,1, CV_32FC1, r).clone(), // red
Mat(64,1, CV_32FC1, g).clone(), // green
Mat(64,1, CV_32FC1, b).clone(), // blue
n); // number of sample points
}
};
// Equals the GNU Octave colormap "hsv".
class HSV : public ColorMap {
public:
HSV() : ColorMap() {
init(256);
}
HSV(int n) : ColorMap() {
init(n);
}
void init(int n) {
float r[] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0.9523809523809526, 0.8571428571428568, 0.7619047619047614, 0.6666666666666665, 0.5714285714285716, 0.4761904761904763, 0.3809523809523805, 0.2857142857142856, 0.1904761904761907, 0.0952380952380949, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.09523809523809557, 0.1904761904761905, 0.2857142857142854, 0.3809523809523809, 0.4761904761904765, 0.5714285714285714, 0.6666666666666663, 0.7619047619047619, 0.8571428571428574, 0.9523809523809523, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
float g[] = { 0, 0.09523809523809523, 0.1904761904761905, 0.2857142857142857, 0.3809523809523809, 0.4761904761904762, 0.5714285714285714, 0.6666666666666666, 0.7619047619047619, 0.8571428571428571, 0.9523809523809523, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0.9523809523809526, 0.8571428571428577, 0.7619047619047619, 0.6666666666666665, 0.5714285714285716, 0.4761904761904767, 0.3809523809523814, 0.2857142857142856, 0.1904761904761907, 0.09523809523809579, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
float b[] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.09523809523809523, 0.1904761904761905, 0.2857142857142857, 0.3809523809523809, 0.4761904761904762, 0.5714285714285714, 0.6666666666666666, 0.7619047619047619, 0.8571428571428571, 0.9523809523809523, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0.9523809523809526, 0.8571428571428577, 0.7619047619047614, 0.6666666666666665, 0.5714285714285716, 0.4761904761904767, 0.3809523809523805, 0.2857142857142856, 0.1904761904761907, 0.09523809523809579, 0};
Mat X = linspace(0,1,64);
this->_lut = ColorMap::linear_colormap(X,
Mat(64,1, CV_32FC1, r).clone(), // red
Mat(64,1, CV_32FC1, g).clone(), // green
Mat(64,1, CV_32FC1, b).clone(), // blue
n); // number of sample points
}
};
// Equals the GNU Octave colormap "pink".
class Pink : public ColorMap {
public:
Pink() : ColorMap() {
init(256);
}
Pink(int n) : ColorMap() {
init(n);
}
void init(int n) {
float r[] = { 0, 0.1571348402636772, 0.2222222222222222, 0.2721655269759087, 0.3142696805273544, 0.3513641844631533, 0.3849001794597505, 0.415739709641549, 0.4444444444444444, 0.4714045207910317, 0.4969039949999532, 0.5211573066470477, 0.5443310539518174, 0.5665577237325317, 0.5879447357921312, 0.6085806194501846, 0.6285393610547089, 0.6478835438717, 0.6666666666666666, 0.6849348892187751, 0.7027283689263065, 0.7200822998230956, 0.7370277311900888, 0.753592220347252, 0.7663560447348133, 0.7732293307186413, 0.7800420555749596, 0.7867957924694432, 0.7934920476158722, 0.8001322641986387, 0.8067178260046388, 0.8132500607904444, 0.8197302434079591, 0.8261595987094034, 0.8325393042503717, 0.8388704928078611, 0.8451542547285166, 0.8513916401208816, 0.8575836609041332, 0.8637312927246217, 0.8698354767504924, 0.8758971213537393, 0.8819171036881968, 0.8878962711712378, 0.8938354428762595, 0.8997354108424372, 0.9055969413076769, 0.9114207758701963, 0.9172076325837248, 0.9229582069908971, 0.9286731730990523, 0.9343531843023135, 0.9399988742535192, 0.9456108576893002, 0.9511897312113418, 0.9567360740266436, 0.9622504486493763, 0.9677334015667416, 0.9731854638710686, 0.9786071518602129, 0.9839989676081821, 0.9893613995077727, 0.9946949227868761, 1};
float g[] = { 0, 0.1028688999747279, 0.1454785934906616, 0.1781741612749496, 0.2057377999494559, 0.2300218531141181, 0.2519763153394848, 0.2721655269759087, 0.2909571869813232, 0.3086066999241838, 0.3253000243161777, 0.3411775438127727, 0.3563483225498992, 0.3708990935094579, 0.3849001794597505, 0.3984095364447979, 0.4114755998989117, 0.4241393401869012, 0.4364357804719847, 0.4483951394230328, 0.4600437062282361, 0.4714045207910317, 0.4824979096371639, 0.4933419132673033, 0.5091750772173156, 0.5328701692569688, 0.5555555555555556, 0.5773502691896257, 0.5983516452371671, 0.6186404847588913, 0.6382847385042254, 0.6573421981221795, 0.6758625033664688, 0.6938886664887108, 0.7114582486036499, 0.7286042804780002, 0.7453559924999299, 0.7617394000445604, 0.7777777777777778, 0.7934920476158723, 0.8089010988089465, 0.8240220541217402, 0.8388704928078611, 0.8534606386520677, 0.8678055195451838, 0.8819171036881968, 0.8958064164776166, 0.9094836413191612, 0.9172076325837248, 0.9229582069908971, 0.9286731730990523, 0.9343531843023135, 0.9399988742535192, 0.9456108576893002, 0.9511897312113418, 0.9567360740266436, 0.9622504486493763, 0.9677334015667416, 0.9731854638710686, 0.9786071518602129, 0.9839989676081821, 0.9893613995077727, 0.9946949227868761, 1};
float b[] = { 0, 0.1028688999747279, 0.1454785934906616, 0.1781741612749496, 0.2057377999494559, 0.2300218531141181, 0.2519763153394848, 0.2721655269759087, 0.2909571869813232, 0.3086066999241838, 0.3253000243161777, 0.3411775438127727, 0.3563483225498992, 0.3708990935094579, 0.3849001794597505, 0.3984095364447979, 0.4114755998989117, 0.4241393401869012, 0.4364357804719847, 0.4483951394230328, 0.4600437062282361, 0.4714045207910317, 0.4824979096371639, 0.4933419132673033, 0.5039526306789697, 0.5143444998736397, 0.5245305283129621, 0.5345224838248488, 0.5443310539518174, 0.5539659798925444, 0.563436169819011, 0.5727497953228163, 0.5819143739626463, 0.5909368402852788, 0.5998236072282915, 0.6085806194501846, 0.6172133998483676, 0.6257270902992705, 0.6341264874742278, 0.642416074439621, 0.6506000486323554, 0.6586823467062358, 0.6666666666666666, 0.6745564876468501, 0.6823550876255453, 0.6900655593423541, 0.6976908246297114, 0.7052336473499384, 0.7237468644557459, 0.7453559924999298, 0.7663560447348133, 0.7867957924694432, 0.8067178260046388, 0.8261595987094034, 0.8451542547285166, 0.8637312927246217, 0.8819171036881968, 0.8997354108424372, 0.9172076325837248, 0.9343531843023135, 0.9511897312113418, 0.9677334015667416, 0.9839989676081821, 1};
Mat X = linspace(0,1,64);
this->_lut = ColorMap::linear_colormap(X,
Mat(64,1, CV_32FC1, r).clone(), // red
Mat(64,1, CV_32FC1, g).clone(), // green
Mat(64,1, CV_32FC1, b).clone(), // blue
n); // number of sample points
}
};
// Equals the GNU Octave colormap "hot".
class Hot : public ColorMap {
public:
Hot() : ColorMap() {
init(256);
}
Hot(int n) : ColorMap() {
init(n);
}
void init(int n) {
float r[] = { 0, 0.03968253968253968, 0.07936507936507936, 0.119047619047619, 0.1587301587301587, 0.1984126984126984, 0.2380952380952381, 0.2777777777777778, 0.3174603174603174, 0.3571428571428571, 0.3968253968253968, 0.4365079365079365, 0.4761904761904762, 0.5158730158730158, 0.5555555555555556, 0.5952380952380952, 0.6349206349206349, 0.6746031746031745, 0.7142857142857142, 0.753968253968254, 0.7936507936507936, 0.8333333333333333, 0.873015873015873, 0.9126984126984127, 0.9523809523809523, 0.992063492063492, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
float g[] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.03174603174603163, 0.0714285714285714, 0.1111111111111112, 0.1507936507936507, 0.1904761904761905, 0.23015873015873, 0.2698412698412698, 0.3095238095238093, 0.3492063492063491, 0.3888888888888888, 0.4285714285714284, 0.4682539682539679, 0.5079365079365079, 0.5476190476190477, 0.5873015873015872, 0.6269841269841268, 0.6666666666666665, 0.7063492063492065, 0.746031746031746, 0.7857142857142856, 0.8253968253968254, 0.8650793650793651, 0.9047619047619047, 0.9444444444444442, 0.984126984126984, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
float b[] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.04761904761904745, 0.1269841269841265, 0.2063492063492056, 0.2857142857142856, 0.3650793650793656, 0.4444444444444446, 0.5238095238095237, 0.6031746031746028, 0.6825396825396828, 0.7619047619047619, 0.8412698412698409, 0.92063492063492, 1};
Mat X = linspace(0,1,64);
this->_lut = ColorMap::linear_colormap(X,
Mat(64,1, CV_32FC1, r).clone(), // red
Mat(64,1, CV_32FC1, g).clone(), // green
Mat(64,1, CV_32FC1, b).clone(), // blue
n); // number of sample points
}
};
void ColorMap::operator()(InputArray _src, OutputArray _dst) const
{
if(_lut.total() != 256)
CV_Error(Error::StsAssert, "cv::LUT only supports tables of size 256.");
Mat src = _src.getMat();
// Return original matrix if wrong type is given (is fail loud better here?)
if(src.type() != CV_8UC1 && src.type() != CV_8UC3)
{
src.copyTo(_dst);
return;
}
// Turn into a BGR matrix into its grayscale representation.
if(src.type() == CV_8UC3)
cvtColor(src.clone(), src, COLOR_BGR2GRAY);
cvtColor(src.clone(), src, COLOR_GRAY2BGR);
// Apply the ColorMap.
LUT(src, _lut, _dst);
}
Mat ColorMap::linear_colormap(InputArray X,
InputArray r, InputArray g, InputArray b,
InputArray xi) {
Mat lut, lut8;
Mat planes[] = {
interp1(X, b, xi),
interp1(X, g, xi),
interp1(X, r, xi)};
merge(planes, 3, lut);
lut.convertTo(lut8, CV_8U, 255.);
return lut8;
}
}
void applyColorMap(InputArray src, OutputArray dst, int colormap)
{
colormap::ColorMap* cm =
colormap == COLORMAP_AUTUMN ? (colormap::ColorMap*)(new colormap::Autumn) :
colormap == COLORMAP_BONE ? (colormap::ColorMap*)(new colormap::Bone) :
colormap == COLORMAP_COOL ? (colormap::ColorMap*)(new colormap::Cool) :
colormap == COLORMAP_HOT ? (colormap::ColorMap*)(new colormap::Hot) :
colormap == COLORMAP_HSV ? (colormap::ColorMap*)(new colormap::HSV) :
colormap == COLORMAP_JET ? (colormap::ColorMap*)(new colormap::Jet) :
colormap == COLORMAP_OCEAN ? (colormap::ColorMap*)(new colormap::Ocean) :
colormap == COLORMAP_PINK ? (colormap::ColorMap*)(new colormap::Pink) :
colormap == COLORMAP_RAINBOW ? (colormap::ColorMap*)(new colormap::Rainbow) :
colormap == COLORMAP_SPRING ? (colormap::ColorMap*)(new colormap::Spring) :
colormap == COLORMAP_SUMMER ? (colormap::ColorMap*)(new colormap::Summer) :
colormap == COLORMAP_WINTER ? (colormap::ColorMap*)(new colormap::Winter) : 0;
if( !cm )
CV_Error( Error::StsBadArg, "Unknown colormap id; use one of COLORMAP_*");
(*cm)(src, dst);
delete cm;
}
}

134
modules/contrib/src/colortracker.cpp
View File

@@ -1,134 +0,0 @@
//*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2008-2011, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of Intel Corporation may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "precomp.hpp"
#include "opencv2/contrib/hybridtracker.hpp"
using namespace cv;
CvMeanShiftTracker::CvMeanShiftTracker(CvMeanShiftTrackerParams _params) : params(_params)
{
}
CvMeanShiftTracker::~CvMeanShiftTracker()
{
}
void CvMeanShiftTracker::newTrackingWindow(Mat image, Rect selection)
{
hist.release();
int channels[] = { 0, 0 , 1, 1};
float hrange[] = { 0, 180 };
float srange[] = { 0, 1 };
const float* ranges[] = {hrange, srange};
cvtColor(image, hsv, COLOR_BGR2HSV);
inRange(hsv, Scalar(0, 30, MIN(10, 256)), Scalar(180, 256, MAX(10, 256)), mask);
hue.create(hsv.size(), CV_8UC2);
mixChannels(&hsv, 1, &hue, 1, channels, 2);
Mat roi(hue, selection);
Mat mskroi(mask, selection);
int ch[] = {0, 1};
int chsize[] = {32, 32};
calcHist(&roi, 1, ch, mskroi, hist, 1, chsize, ranges);
normalize(hist, hist, 0, 255, CV_MINMAX);
prev_trackwindow = selection;
}
RotatedRect CvMeanShiftTracker::updateTrackingWindow(Mat image)
{
int channels[] = { 0, 0 , 1, 1};
float hrange[] = { 0, 180 };
float srange[] = { 0, 1 };
const float* ranges[] = {hrange, srange};
cvtColor(image, hsv, COLOR_BGR2HSV);
inRange(hsv, Scalar(0, 30, MIN(10, 256)), Scalar(180, 256, MAX(10, 256)), mask);
hue.create(hsv.size(), CV_8UC2);
mixChannels(&hsv, 1, &hue, 1, channels, 2);
int ch[] = {0, 1};
calcBackProject(&hue, 1, ch, hist, backproj, ranges);
backproj &= mask;
prev_trackbox = CamShift(backproj, prev_trackwindow, TermCriteria(CV_TERMCRIT_EPS | CV_TERMCRIT_ITER, 10, 1));
int cols = backproj.cols, rows = backproj.rows, r = (MIN(cols, rows) + 5) / 6;
prev_trackwindow = Rect(prev_trackwindow.x - r, prev_trackwindow.y - r, prev_trackwindow.x + r,
prev_trackwindow.y + r) & Rect(0, 0, cols, rows);
prev_center.x = (float)(prev_trackwindow.x + prev_trackwindow.width / 2);
prev_center.y = (float)(prev_trackwindow.y + prev_trackwindow.height / 2);
#ifdef DEBUG_HYTRACKER
ellipse(image, prev_trackbox, Scalar(0, 0, 255), 1, CV_AA);
#endif
return prev_trackbox;
}
Mat CvMeanShiftTracker::getHistogramProjection(int type)
{
Mat ms_backproj_f(backproj.size(), type);
backproj.convertTo(ms_backproj_f, type);
return ms_backproj_f;
}
void CvMeanShiftTracker::setTrackingWindow(Rect window)
{
prev_trackwindow = window;
}
Rect CvMeanShiftTracker::getTrackingWindow()
{
return prev_trackwindow;
}
RotatedRect CvMeanShiftTracker::getTrackingEllipse()
{
return prev_trackbox;
}
Point2f CvMeanShiftTracker::getTrackingCenter()
{
return prev_center;
}

901
modules/contrib/src/facerec.cpp
View File

@@ -1,901 +0,0 @@
/*
* Copyright (c) 2011,2012. Philipp Wagner <bytefish[at]gmx[dot]de>.
* Released to public domain under terms of the BSD Simplified license.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of the organization nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* See <http://www.opensource.org/licenses/bsd-license>
*/
#include "precomp.hpp"
#include <set>
#include <limits>
namespace cv
{
// Reads a sequence from a FileNode::SEQ with type _Tp into a result vector.
template<typename _Tp>
inline void readFileNodeList(const FileNode& fn, std::vector<_Tp>& result) {
if (fn.type() == FileNode::SEQ) {
for (FileNodeIterator it = fn.begin(); it != fn.end();) {
_Tp item;
it >> item;
result.push_back(item);
}
}
}
// Writes the a list of given items to a cv::FileStorage.
template<typename _Tp>
inline void writeFileNodeList(FileStorage& fs, const String& name,
const std::vector<_Tp>& items) {
// typedefs
typedef typename std::vector<_Tp>::const_iterator constVecIterator;
// write the elements in item to fs
fs << name << "[";
for (constVecIterator it = items.begin(); it != items.end(); ++it) {
fs << *it;
}
fs << "]";
}
static Mat asRowMatrix(InputArrayOfArrays src, int rtype, double alpha=1, double beta=0) {
// make sure the input data is a vector of matrices or vector of vector
if(src.kind() != _InputArray::STD_VECTOR_MAT && src.kind() != _InputArray::STD_VECTOR_VECTOR) {
String error_message = "The data is expected as InputArray::STD_VECTOR_MAT (a std::vector<Mat>) or _InputArray::STD_VECTOR_VECTOR (a std::vector< std::vector<...> >).";
CV_Error(Error::StsBadArg, error_message);
}
// number of samples
size_t n = src.total();
// return empty matrix if no matrices given
if(n == 0)
return Mat();
// dimensionality of (reshaped) samples
size_t d = src.getMat(0).total();
// create data matrix
Mat data((int)n, (int)d, rtype);
// now copy data
for(unsigned int i = 0; i < n; i++) {
// make sure data can be reshaped, throw exception if not!
if(src.getMat(i).total() != d) {
String error_message = format("Wrong number of elements in matrix #%d! Expected %d was %d.", i, d, src.getMat(i).total());
CV_Error(Error::StsBadArg, error_message);
}
// get a hold of the current row
Mat xi = data.row(i);
// make reshape happy by cloning for non-continuous matrices
if(src.getMat(i).isContinuous()) {
src.getMat(i).reshape(1, 1).convertTo(xi, rtype, alpha, beta);
} else {
src.getMat(i).clone().reshape(1, 1).convertTo(xi, rtype, alpha, beta);
}
}
return data;
}
// Removes duplicate elements in a given vector.
template<typename _Tp>
inline std::vector<_Tp> remove_dups(const std::vector<_Tp>& src) {
typedef typename std::set<_Tp>::const_iterator constSetIterator;
typedef typename std::vector<_Tp>::const_iterator constVecIterator;
std::set<_Tp> set_elems;
for (constVecIterator it = src.begin(); it != src.end(); ++it)
set_elems.insert(*it);
std::vector<_Tp> elems;
for (constSetIterator it = set_elems.begin(); it != set_elems.end(); ++it)
elems.push_back(*it);
return elems;
}
// Turk, M., and Pentland, A. "Eigenfaces for recognition.". Journal of
// Cognitive Neuroscience 3 (1991), 71–86.
class Eigenfaces : public FaceRecognizer
{
private:
int _num_components;
double _threshold;
std::vector<Mat> _projections;
Mat _labels;
Mat _eigenvectors;
Mat _eigenvalues;
Mat _mean;
public:
using FaceRecognizer::save;
using FaceRecognizer::load;
// Initializes an empty Eigenfaces model.
Eigenfaces(int num_components = 0, double threshold = DBL_MAX) :
_num_components(num_components),
_threshold(threshold) {}
// Initializes and computes an Eigenfaces model with images in src and
// corresponding labels in labels. num_components will be kept for
// classification.
Eigenfaces(InputArrayOfArrays src, InputArray labels,
int num_components = 0, double threshold = DBL_MAX) :
_num_components(num_components),
_threshold(threshold) {
train(src, labels);
}
// Computes an Eigenfaces model with images in src and corresponding labels
// in labels.
void train(InputArrayOfArrays src, InputArray labels);
// Predicts the label of a query image in src.
int predict(InputArray src) const;
// Predicts the label and confidence for a given sample.
void predict(InputArray _src, int &label, double &dist) const;
// See FaceRecognizer::load.
void load(const FileStorage& fs);
// See FaceRecognizer::save.
void save(FileStorage& fs) const;
AlgorithmInfo* info() const;
};
// Belhumeur, P. N., Hespanha, J., and Kriegman, D. "Eigenfaces vs. Fisher-
// faces: Recognition using class specific linear projection.". IEEE
// Transactions on Pattern Analysis and Machine Intelligence 19, 7 (1997),
// 711–720.
class Fisherfaces: public FaceRecognizer
{
private:
int _num_components;
double _threshold;
Mat _eigenvectors;
Mat _eigenvalues;
Mat _mean;
std::vector<Mat> _projections;
Mat _labels;
public:
using FaceRecognizer::save;
using FaceRecognizer::load;
// Initializes an empty Fisherfaces model.
Fisherfaces(int num_components = 0, double threshold = DBL_MAX) :
_num_components(num_components),
_threshold(threshold) {}
// Initializes and computes a Fisherfaces model with images in src and
// corresponding labels in labels. num_components will be kept for
// classification.
Fisherfaces(InputArrayOfArrays src, InputArray labels,
int num_components = 0, double threshold = DBL_MAX) :
_num_components(num_components),
_threshold(threshold) {
train(src, labels);
}
~Fisherfaces() {}
// Computes a Fisherfaces model with images in src and corresponding labels
// in labels.
void train(InputArrayOfArrays src, InputArray labels);
// Predicts the label of a query image in src.
int predict(InputArray src) const;
// Predicts the label and confidence for a given sample.
void predict(InputArray _src, int &label, double &dist) const;
// See FaceRecognizer::load.
void load(const FileStorage& fs);
// See FaceRecognizer::save.
void save(FileStorage& fs) const;
AlgorithmInfo* info() const;
};
// Face Recognition based on Local Binary Patterns.
//
// Ahonen T, Hadid A. and Pietikäinen M. "Face description with local binary
// patterns: Application to face recognition." IEEE Transactions on Pattern
// Analysis and Machine Intelligence, 28(12):2037-2041.
//
class LBPH : public FaceRecognizer
{
private:
int _grid_x;
int _grid_y;
int _radius;
int _neighbors;
double _threshold;
std::vector<Mat> _histograms;
Mat _labels;
// Computes a LBPH model with images in src and
// corresponding labels in labels, possibly preserving
// old model data.
void train(InputArrayOfArrays src, InputArray labels, bool preserveData);
public:
using FaceRecognizer::save;
using FaceRecognizer::load;
// Initializes this LBPH Model. The current implementation is rather fixed
// as it uses the Extended Local Binary Patterns per default.
//
// radius, neighbors are used in the local binary patterns creation.
// grid_x, grid_y control the grid size of the spatial histograms.
LBPH(int radius_=1, int neighbors_=8,
int gridx=8, int gridy=8,
double threshold = DBL_MAX) :
_grid_x(gridx),
_grid_y(gridy),
_radius(radius_),
_neighbors(neighbors_),
_threshold(threshold) {}
// Initializes and computes this LBPH Model. The current implementation is
// rather fixed as it uses the Extended Local Binary Patterns per default.
//
// (radius=1), (neighbors=8) are used in the local binary patterns creation.
// (grid_x=8), (grid_y=8) controls the grid size of the spatial histograms.
LBPH(InputArrayOfArrays src,
InputArray labels,
int radius_=1, int neighbors_=8,
int gridx=8, int gridy=8,
double threshold = DBL_MAX) :
_grid_x(gridx),
_grid_y(gridy),
_radius(radius_),
_neighbors(neighbors_),
_threshold(threshold) {
train(src, labels);
}
~LBPH() { }
// Computes a LBPH model with images in src and
// corresponding labels in labels.
void train(InputArrayOfArrays src, InputArray labels);
// Updates this LBPH model with images in src and
// corresponding labels in labels.
void update(InputArrayOfArrays src, InputArray labels);
// Predicts the label of a query image in src.
int predict(InputArray src) const;
// Predicts the label and confidence for a given sample.
void predict(InputArray _src, int &label, double &dist) const;
// See FaceRecognizer::load.
void load(const FileStorage& fs);
// See FaceRecognizer::save.
void save(FileStorage& fs) const;
// Getter functions.
int neighbors() const { return _neighbors; }
int radius() const { return _radius; }
int grid_x() const { return _grid_x; }
int grid_y() const { return _grid_y; }
AlgorithmInfo* info() const;
};
//------------------------------------------------------------------------------
// FaceRecognizer
//------------------------------------------------------------------------------
void FaceRecognizer::update(InputArrayOfArrays src, InputArray labels ) {
if( dynamic_cast<LBPH*>(this) != 0 )
{
dynamic_cast<LBPH*>(this)->update( src, labels );
return;
}
String error_msg = format("This FaceRecognizer (%s) does not support updating, you have to use FaceRecognizer::train to update it.", this->name().c_str());
CV_Error(Error::StsNotImplemented, error_msg);
}
void FaceRecognizer::save(const String& filename) const {
FileStorage fs(filename, FileStorage::WRITE);
if (!fs.isOpened())
CV_Error(Error::StsError, "File can't be opened for writing!");
this->save(fs);
fs.release();
}
void FaceRecognizer::load(const String& filename) {
FileStorage fs(filename, FileStorage::READ);
if (!fs.isOpened())
CV_Error(Error::StsError, "File can't be opened for writing!");
this->load(fs);
fs.release();
}
//------------------------------------------------------------------------------
// Eigenfaces
//------------------------------------------------------------------------------
void Eigenfaces::train(InputArrayOfArrays _src, InputArray _local_labels) {
if(_src.total() == 0) {
String error_message = format("Empty training data was given. You'll need more than one sample to learn a model.");
CV_Error(Error::StsBadArg, error_message);
} else if(_local_labels.getMat().type() != CV_32SC1) {
String error_message = format("Labels must be given as integer (CV_32SC1). Expected %d, but was %d.", CV_32SC1, _local_labels.type());
CV_Error(Error::StsBadArg, error_message);
}
// make sure data has correct size
if(_src.total() > 1) {
for(int i = 1; i < static_cast<int>(_src.total()); i++) {
if(_src.getMat(i-1).total() != _src.getMat(i).total()) {
String error_message = format("In the Eigenfaces method all input samples (training images) must be of equal size! Expected %d pixels, but was %d pixels.", _src.getMat(i-1).total(), _src.getMat(i).total());
CV_Error(Error::StsUnsupportedFormat, error_message);
}
}
}
// get labels
Mat labels = _local_labels.getMat();
// observations in row
Mat data = asRowMatrix(_src, CV_64FC1);
// number of samples
int n = data.rows;
// assert there are as much samples as labels
if(static_cast<int>(labels.total()) != n) {
String error_message = format("The number of samples (src) must equal the number of labels (labels)! len(src)=%d, len(labels)=%d.", n, labels.total());
CV_Error(Error::StsBadArg, error_message);
}
// clear existing model data
_labels.release();
_projections.clear();
// clip number of components to be valid
if((_num_components <= 0) || (_num_components > n))
_num_components = n;
// perform the PCA
PCA pca(data, Mat(), PCA::DATA_AS_ROW, _num_components);
// copy the PCA results
_mean = pca.mean.reshape(1,1); // store the mean vector
_eigenvalues = pca.eigenvalues.clone(); // eigenvalues by row
transpose(pca.eigenvectors, _eigenvectors); // eigenvectors by column
// store labels for prediction
_labels = labels.clone();
// save projections
for(int sampleIdx = 0; sampleIdx < data.rows; sampleIdx++) {
Mat p = subspaceProject(_eigenvectors, _mean, data.row(sampleIdx));
_projections.push_back(p);
}
}
void Eigenfaces::predict(InputArray _src, int &minClass, double &minDist) const {
// get data
Mat src = _src.getMat();
// make sure the user is passing correct data
if(_projections.empty()) {
// throw error if no data (or simply return -1?)
String error_message = "This Eigenfaces model is not computed yet. Did you call Eigenfaces::train?";
CV_Error(Error::StsError, error_message);
} else if(_eigenvectors.rows != static_cast<int>(src.total())) {
// check data alignment just for clearer exception messages
String error_message = format("Wrong input image size. Reason: Training and Test images must be of equal size! Expected an image with %d elements, but got %d.", _eigenvectors.rows, src.total());
CV_Error(Error::StsBadArg, error_message);
}
// project into PCA subspace
Mat q = subspaceProject(_eigenvectors, _mean, src.reshape(1,1));
minDist = DBL_MAX;
minClass = -1;
for(size_t sampleIdx = 0; sampleIdx < _projections.size(); sampleIdx++) {
double dist = norm(_projections[sampleIdx], q, NORM_L2);
if((dist < minDist) && (dist < _threshold)) {
minDist = dist;
minClass = _labels.at<int>((int)sampleIdx);
}
}
}
int Eigenfaces::predict(InputArray _src) const {
int label;
double dummy;
predict(_src, label, dummy);
return label;
}
void Eigenfaces::load(const FileStorage& fs) {
//read matrices
fs["num_components"] >> _num_components;
fs["mean"] >> _mean;
fs["eigenvalues"] >> _eigenvalues;
fs["eigenvectors"] >> _eigenvectors;
// read sequences
readFileNodeList(fs["projections"], _projections);
fs["labels"] >> _labels;
}
void Eigenfaces::save(FileStorage& fs) const {
// write matrices
fs << "num_components" << _num_components;
fs << "mean" << _mean;
fs << "eigenvalues" << _eigenvalues;
fs << "eigenvectors" << _eigenvectors;
// write sequences
writeFileNodeList(fs, "projections", _projections);
fs << "labels" << _labels;
}
//------------------------------------------------------------------------------
// Fisherfaces
//------------------------------------------------------------------------------
void Fisherfaces::train(InputArrayOfArrays src, InputArray _lbls) {
if(src.total() == 0) {
String error_message = format("Empty training data was given. You'll need more than one sample to learn a model.");
CV_Error(Error::StsBadArg, error_message);
} else if(_lbls.getMat().type() != CV_32SC1) {
String error_message = format("Labels must be given as integer (CV_32SC1). Expected %d, but was %d.", CV_32SC1, _lbls.type());
CV_Error(Error::StsBadArg, error_message);
}
// make sure data has correct size
if(src.total() > 1) {
for(int i = 1; i < static_cast<int>(src.total()); i++) {
if(src.getMat(i-1).total() != src.getMat(i).total()) {
String error_message = format("In the Fisherfaces method all input samples (training images) must be of equal size! Expected %d pixels, but was %d pixels.", src.getMat(i-1).total(), src.getMat(i).total());
CV_Error(Error::StsUnsupportedFormat, error_message);
}
}
}
// get data
Mat labels = _lbls.getMat();
Mat data = asRowMatrix(src, CV_64FC1);
// number of samples
int N = data.rows;
// make sure labels are passed in correct shape
if(labels.total() != (size_t) N) {
String error_message = format("The number of samples (src) must equal the number of labels (labels)! len(src)=%d, len(labels)=%d.", N, labels.total());
CV_Error(Error::StsBadArg, error_message);
} else if(labels.rows != 1 && labels.cols != 1) {
String error_message = format("Expected the labels in a matrix with one row or column! Given dimensions are rows=%s, cols=%d.", labels.rows, labels.cols);
CV_Error(Error::StsBadArg, error_message);
}
// clear existing model data
_labels.release();
_projections.clear();
// safely copy from cv::Mat to std::vector
std::vector<int> ll;
for(unsigned int i = 0; i < labels.total(); i++) {
ll.push_back(labels.at<int>(i));
}
// get the number of unique classes
int C = (int) remove_dups(ll).size();
// clip number of components to be a valid number
if((_num_components <= 0) || (_num_components > (C-1)))
_num_components = (C-1);
// perform a PCA and keep (N-C) components
PCA pca(data, Mat(), PCA::DATA_AS_ROW, (N-C));
// project the data and perform a LDA on it
LDA lda(pca.project(data),labels, _num_components);
// store the total mean vector
_mean = pca.mean.reshape(1,1);
// store labels
_labels = labels.clone();
// store the eigenvalues of the discriminants
lda.eigenvalues().convertTo(_eigenvalues, CV_64FC1);
// Now calculate the projection matrix as pca.eigenvectors * lda.eigenvectors.
// Note: OpenCV stores the eigenvectors by row, so we need to transpose it!
gemm(pca.eigenvectors, lda.eigenvectors(), 1.0, Mat(), 0.0, _eigenvectors, GEMM_1_T);
// store the projections of the original data
for(int sampleIdx = 0; sampleIdx < data.rows; sampleIdx++) {
Mat p = subspaceProject(_eigenvectors, _mean, data.row(sampleIdx));
_projections.push_back(p);
}
}
void Fisherfaces::predict(InputArray _src, int &minClass, double &minDist) const {
Mat src = _src.getMat();
// check data alignment just for clearer exception messages
if(_projections.empty()) {
// throw error if no data (or simply return -1?)
String error_message = "This Fisherfaces model is not computed yet. Did you call Fisherfaces::train?";
CV_Error(Error::StsBadArg, error_message);
} else if(src.total() != (size_t) _eigenvectors.rows) {
String error_message = format("Wrong input image size. Reason: Training and Test images must be of equal size! Expected an image with %d elements, but got %d.", _eigenvectors.rows, src.total());
CV_Error(Error::StsBadArg, error_message);
}
// project into LDA subspace
Mat q = subspaceProject(_eigenvectors, _mean, src.reshape(1,1));
// find 1-nearest neighbor
minDist = DBL_MAX;
minClass = -1;
for(size_t sampleIdx = 0; sampleIdx < _projections.size(); sampleIdx++) {
double dist = norm(_projections[sampleIdx], q, NORM_L2);
if((dist < minDist) && (dist < _threshold)) {
minDist = dist;
minClass = _labels.at<int>((int)sampleIdx);
}
}
}
int Fisherfaces::predict(InputArray _src) const {
int label;
double dummy;
predict(_src, label, dummy);
return label;
}
// See FaceRecognizer::load.
void Fisherfaces::load(const FileStorage& fs) {
//read matrices
fs["num_components"] >> _num_components;
fs["mean"] >> _mean;
fs["eigenvalues"] >> _eigenvalues;
fs["eigenvectors"] >> _eigenvectors;
// read sequences
readFileNodeList(fs["projections"], _projections);
fs["labels"] >> _labels;
}
// See FaceRecognizer::save.
void Fisherfaces::save(FileStorage& fs) const {
// write matrices
fs << "num_components" << _num_components;
fs << "mean" << _mean;
fs << "eigenvalues" << _eigenvalues;
fs << "eigenvectors" << _eigenvectors;
// write sequences
writeFileNodeList(fs, "projections", _projections);
fs << "labels" << _labels;
}
//------------------------------------------------------------------------------
// LBPH
//------------------------------------------------------------------------------
template <typename _Tp> static
void olbp_(InputArray _src, OutputArray _dst) {
// get matrices
Mat src = _src.getMat();
// allocate memory for result
_dst.create(src.rows-2, src.cols-2, CV_8UC1);
Mat dst = _dst.getMat();
// zero the result matrix
dst.setTo(0);
// calculate patterns
for(int i=1;i<src.rows-1;i++) {
for(int j=1;j<src.cols-1;j++) {
_Tp center = src.at<_Tp>(i,j);
unsigned char code = 0;
code |= (src.at<_Tp>(i-1,j-1) >= center) << 7;
code |= (src.at<_Tp>(i-1,j) >= center) << 6;
code |= (src.at<_Tp>(i-1,j+1) >= center) << 5;
code |= (src.at<_Tp>(i,j+1) >= center) << 4;
code |= (src.at<_Tp>(i+1,j+1) >= center) << 3;
code |= (src.at<_Tp>(i+1,j) >= center) << 2;
code |= (src.at<_Tp>(i+1,j-1) >= center) << 1;
code |= (src.at<_Tp>(i,j-1) >= center) << 0;
dst.at<unsigned char>(i-1,j-1) = code;
}
}
}
//------------------------------------------------------------------------------
// cv::elbp
//------------------------------------------------------------------------------
template <typename _Tp> static
inline void elbp_(InputArray _src, OutputArray _dst, int radius, int neighbors) {
//get matrices
Mat src = _src.getMat();
// allocate memory for result
_dst.create(src.rows-2*radius, src.cols-2*radius, CV_32SC1);
Mat dst = _dst.getMat();
// zero
dst.setTo(0);
for(int n=0; n<neighbors; n++) {
// sample points
float x = static_cast<float>(radius * cos(2.0*CV_PI*n/static_cast<float>(neighbors)));
float y = static_cast<float>(-radius * sin(2.0*CV_PI*n/static_cast<float>(neighbors)));
// relative indices
int fx = static_cast<int>(floor(x));
int fy = static_cast<int>(floor(y));
int cx = static_cast<int>(ceil(x));
int cy = static_cast<int>(ceil(y));
// fractional part
float ty = y - fy;
float tx = x - fx;
// set interpolation weights
float w1 = (1 - tx) * (1 - ty);
float w2 = tx * (1 - ty);
float w3 = (1 - tx) * ty;
float w4 = tx * ty;
// iterate through your data
for(int i=radius; i < src.rows-radius;i++) {
for(int j=radius;j < src.cols-radius;j++) {
// calculate interpolated value
float t = static_cast<float>(w1*src.at<_Tp>(i+fy,j+fx) + w2*src.at<_Tp>(i+fy,j+cx) + w3*src.at<_Tp>(i+cy,j+fx) + w4*src.at<_Tp>(i+cy,j+cx));
// floating point precision, so check some machine-dependent epsilon
dst.at<int>(i-radius,j-radius) += ((t > src.at<_Tp>(i,j)) || (std::abs(t-src.at<_Tp>(i,j)) < std::numeric_limits<float>::epsilon())) << n;
}
}
}
}
static void elbp(InputArray src, OutputArray dst, int radius, int neighbors)
{
int type = src.type();
switch (type) {
case CV_8SC1: elbp_<char>(src,dst, radius, neighbors); break;
case CV_8UC1: elbp_<unsigned char>(src, dst, radius, neighbors); break;
case CV_16SC1: elbp_<short>(src,dst, radius, neighbors); break;
case CV_16UC1: elbp_<unsigned short>(src,dst, radius, neighbors); break;
case CV_32SC1: elbp_<int>(src,dst, radius, neighbors); break;
case CV_32FC1: elbp_<float>(src,dst, radius, neighbors); break;
case CV_64FC1: elbp_<double>(src,dst, radius, neighbors); break;
default:
String error_msg = format("Using Original Local Binary Patterns for feature extraction only works on single-channel images (given %d). Please pass the image data as a grayscale image!", type);
CV_Error(Error::StsNotImplemented, error_msg);
break;
}
}
static Mat
histc_(const Mat& src, int minVal=0, int maxVal=255, bool normed=false)
{
Mat result;
// Establish the number of bins.
int histSize = maxVal-minVal+1;
// Set the ranges.
float range[] = { static_cast<float>(minVal), static_cast<float>(maxVal+1) };
const float* histRange = { range };
// calc histogram
calcHist(&src, 1, 0, Mat(), result, 1, &histSize, &histRange, true, false);
// normalize
if(normed) {
result /= (int)src.total();
}
return result.reshape(1,1);
}
static Mat histc(InputArray _src, int minVal, int maxVal, bool normed)
{
Mat src = _src.getMat();
switch (src.type()) {
case CV_8SC1:
return histc_(Mat_<float>(src), minVal, maxVal, normed);
break;
case CV_8UC1:
return histc_(src, minVal, maxVal, normed);
break;
case CV_16SC1:
return histc_(Mat_<float>(src), minVal, maxVal, normed);
break;
case CV_16UC1:
return histc_(src, minVal, maxVal, normed);
break;
case CV_32SC1:
return histc_(Mat_<float>(src), minVal, maxVal, normed);
break;
case CV_32FC1:
return histc_(src, minVal, maxVal, normed);
break;
default:
CV_Error(Error::StsUnmatchedFormats, "This type is not implemented yet."); break;
}
return Mat();
}
static Mat spatial_histogram(InputArray _src, int numPatterns,
int grid_x, int grid_y, bool /*normed*/)
{
Mat src = _src.getMat();
// calculate LBP patch size
int width = src.cols/grid_x;
int height = src.rows/grid_y;
// allocate memory for the spatial histogram
Mat result = Mat::zeros(grid_x * grid_y, numPatterns, CV_32FC1);
// return matrix with zeros if no data was given
if(src.empty())
return result.reshape(1,1);
// initial result_row
int resultRowIdx = 0;
// iterate through grid
for(int i = 0; i < grid_y; i++) {
for(int j = 0; j < grid_x; j++) {
Mat src_cell = Mat(src, Range(i*height,(i+1)*height), Range(j*width,(j+1)*width));
Mat cell_hist = histc(src_cell, 0, (numPatterns-1), true);
// copy to the result matrix
Mat result_row = result.row(resultRowIdx);
cell_hist.reshape(1,1).convertTo(result_row, CV_32FC1);
// increase row count in result matrix
resultRowIdx++;
}
}
// return result as reshaped feature vector
return result.reshape(1,1);
}
//------------------------------------------------------------------------------
// wrapper to cv::elbp (extended local binary patterns)
//------------------------------------------------------------------------------
static Mat elbp(InputArray src, int radius, int neighbors) {
Mat dst;
elbp(src, dst, radius, neighbors);
return dst;
}
void LBPH::load(const FileStorage& fs) {
fs["radius"] >> _radius;
fs["neighbors"] >> _neighbors;
fs["grid_x"] >> _grid_x;
fs["grid_y"] >> _grid_y;
//read matrices
readFileNodeList(fs["histograms"], _histograms);
fs["labels"] >> _labels;
}
// See FaceRecognizer::save.
void LBPH::save(FileStorage& fs) const {
fs << "radius" << _radius;
fs << "neighbors" << _neighbors;
fs << "grid_x" << _grid_x;
fs << "grid_y" << _grid_y;
// write matrices
writeFileNodeList(fs, "histograms", _histograms);
fs << "labels" << _labels;
}
void LBPH::train(InputArrayOfArrays _in_src, InputArray _in_labels) {
this->train(_in_src, _in_labels, false);
}
void LBPH::update(InputArrayOfArrays _in_src, InputArray _in_labels) {
// got no data, just return
if(_in_src.total() == 0)
return;
this->train(_in_src, _in_labels, true);
}
void LBPH::train(InputArrayOfArrays _in_src, InputArray _in_labels, bool preserveData) {
if(_in_src.kind() != _InputArray::STD_VECTOR_MAT && _in_src.kind() != _InputArray::STD_VECTOR_VECTOR) {
String error_message = "The images are expected as InputArray::STD_VECTOR_MAT (a std::vector<Mat>) or _InputArray::STD_VECTOR_VECTOR (a std::vector< std::vector<...> >).";
CV_Error(Error::StsBadArg, error_message);
}
if(_in_src.total() == 0) {
String error_message = format("Empty training data was given. You'll need more than one sample to learn a model.");
CV_Error(Error::StsUnsupportedFormat, error_message);
} else if(_in_labels.getMat().type() != CV_32SC1) {
String error_message = format("Labels must be given as integer (CV_32SC1). Expected %d, but was %d.", CV_32SC1, _in_labels.type());
CV_Error(Error::StsUnsupportedFormat, error_message);
}
// get the vector of matrices
std::vector<Mat> src;
_in_src.getMatVector(src);
// get the label matrix
Mat labels = _in_labels.getMat();
// check if data is well- aligned
if(labels.total() != src.size()) {
String error_message = format("The number of samples (src) must equal the number of labels (labels). Was len(samples)=%d, len(labels)=%d.", src.size(), _labels.total());
CV_Error(Error::StsBadArg, error_message);
}
// if this model should be trained without preserving old data, delete old model data
if(!preserveData) {
_labels.release();
_histograms.clear();
}
// append labels to _labels matrix
for(size_t labelIdx = 0; labelIdx < labels.total(); labelIdx++) {
_labels.push_back(labels.at<int>((int)labelIdx));
}
// store the spatial histograms of the original data
for(size_t sampleIdx = 0; sampleIdx < src.size(); sampleIdx++) {
// calculate lbp image
Mat lbp_image = elbp(src[sampleIdx], _radius, _neighbors);
// get spatial histogram from this lbp image
Mat p = spatial_histogram(
lbp_image, /* lbp_image */
static_cast<int>(std::pow(2.0, static_cast<double>(_neighbors))), /* number of possible patterns */
_grid_x, /* grid size x */
_grid_y, /* grid size y */
true);
// add to templates
_histograms.push_back(p);
}
}
void LBPH::predict(InputArray _src, int &minClass, double &minDist) const {
if(_histograms.empty()) {
// throw error if no data (or simply return -1?)
String error_message = "This LBPH model is not computed yet. Did you call the train method?";
CV_Error(Error::StsBadArg, error_message);
}
Mat src = _src.getMat();
// get the spatial histogram from input image
Mat lbp_image = elbp(src, _radius, _neighbors);
Mat query = spatial_histogram(
lbp_image, /* lbp_image */
static_cast<int>(std::pow(2.0, static_cast<double>(_neighbors))), /* number of possible patterns */
_grid_x, /* grid size x */
_grid_y, /* grid size y */
true /* normed histograms */);
// find 1-nearest neighbor
minDist = DBL_MAX;
minClass = -1;
for(size_t sampleIdx = 0; sampleIdx < _histograms.size(); sampleIdx++) {
double dist = compareHist(_histograms[sampleIdx], query, HISTCMP_CHISQR_ALT);
if((dist < minDist) && (dist < _threshold)) {
minDist = dist;
minClass = _labels.at<int>((int) sampleIdx);
}
}
}
int LBPH::predict(InputArray _src) const {
int label;
double dummy;
predict(_src, label, dummy);
return label;
}
Ptr<FaceRecognizer> createEigenFaceRecognizer(int num_components, double threshold)
{
return makePtr<Eigenfaces>(num_components, threshold);
}
Ptr<FaceRecognizer> createFisherFaceRecognizer(int num_components, double threshold)
{
return makePtr<Fisherfaces>(num_components, threshold);
}
Ptr<FaceRecognizer> createLBPHFaceRecognizer(int radius, int neighbors,
int grid_x, int grid_y, double threshold)
{
return makePtr<LBPH>(radius, neighbors, grid_x, grid_y, threshold);
}
CV_INIT_ALGORITHM(Eigenfaces, "FaceRecognizer.Eigenfaces",
obj.info()->addParam(obj, "ncomponents", obj._num_components);
obj.info()->addParam(obj, "threshold", obj._threshold);
obj.info()->addParam(obj, "projections", obj._projections, true);
obj.info()->addParam(obj, "labels", obj._labels, true);
obj.info()->addParam(obj, "eigenvectors", obj._eigenvectors, true);
obj.info()->addParam(obj, "eigenvalues", obj._eigenvalues, true);
obj.info()->addParam(obj, "mean", obj._mean, true))
CV_INIT_ALGORITHM(Fisherfaces, "FaceRecognizer.Fisherfaces",
obj.info()->addParam(obj, "ncomponents", obj._num_components);
obj.info()->addParam(obj, "threshold", obj._threshold);
obj.info()->addParam(obj, "projections", obj._projections, true);
obj.info()->addParam(obj, "labels", obj._labels, true);
obj.info()->addParam(obj, "eigenvectors", obj._eigenvectors, true);
obj.info()->addParam(obj, "eigenvalues", obj._eigenvalues, true);
obj.info()->addParam(obj, "mean", obj._mean, true))
CV_INIT_ALGORITHM(LBPH, "FaceRecognizer.LBPH",
obj.info()->addParam(obj, "radius", obj._radius);
obj.info()->addParam(obj, "neighbors", obj._neighbors);
obj.info()->addParam(obj, "grid_x", obj._grid_x);
obj.info()->addParam(obj, "grid_y", obj._grid_y);
obj.info()->addParam(obj, "threshold", obj._threshold);
obj.info()->addParam(obj, "histograms", obj._histograms, true);
obj.info()->addParam(obj, "labels", obj._labels, true))
bool initModule_contrib()
{
Ptr<Algorithm> efaces = createEigenfaces_ptr_hidden(), ffaces = createFisherfaces_ptr_hidden(), lbph = createLBPH_ptr_hidden();
return efaces->info() != 0 && ffaces->info() != 0 && lbph->info() != 0;
}
}

229
modules/contrib/src/featuretracker.cpp
View File

@@ -1,229 +0,0 @@
//*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2008-2011, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of Intel Corporation may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "precomp.hpp"
#include <stdio.h>
#include <iostream>
#include "opencv2/calib3d.hpp"
#include "opencv2/contrib/hybridtracker.hpp"
#ifdef HAVE_OPENCV_NONFREE
#include "opencv2/nonfree/nonfree.hpp"
static bool makeUseOfNonfree = initModule_nonfree();
#endif
using namespace cv;
CvFeatureTracker::CvFeatureTracker(CvFeatureTrackerParams _params) :
params(_params)
{
switch (params.feature_type)
{
case CvFeatureTrackerParams::SIFT:
dd = Algorithm::create<Feature2D>("Feature2D.SIFT");
if( !dd )
CV_Error(CV_StsNotImplemented, "OpenCV has been compiled without SIFT support");
dd->set("nOctaveLayers", 5);
dd->set("contrastThreshold", 0.04);
dd->set("edgeThreshold", 10.7);
break;
case CvFeatureTrackerParams::SURF:
dd = Algorithm::create<Feature2D>("Feature2D.SURF");
if( !dd )
CV_Error(CV_StsNotImplemented, "OpenCV has been compiled without SURF support");
dd->set("hessianThreshold", 400);
dd->set("nOctaves", 3);
dd->set("nOctaveLayers", 4);
break;
default:
CV_Error(CV_StsBadArg, "Unknown feature type");
break;
}
matcher = makePtr<BFMatcher>(int(NORM_L2));
}
CvFeatureTracker::~CvFeatureTracker()
{
}
void CvFeatureTracker::newTrackingWindow(Mat image, Rect selection)
{
image.copyTo(prev_image);
cvtColor(prev_image, prev_image_bw, COLOR_BGR2GRAY);
prev_trackwindow = selection;
prev_center.x = selection.x;
prev_center.y = selection.y;
ittr = 0;
}
Rect CvFeatureTracker::updateTrackingWindow(Mat image)
{
if(params.feature_type == CvFeatureTrackerParams::OPTICAL_FLOW)
return updateTrackingWindowWithFlow(image);
else
return updateTrackingWindowWithSIFT(image);
}
Rect CvFeatureTracker::updateTrackingWindowWithSIFT(Mat image)
{
ittr++;
std::vector<KeyPoint> prev_keypoints, curr_keypoints;
std::vector<Point2f> prev_keys, curr_keys;
Mat prev_desc, curr_desc;
Rect window = prev_trackwindow;
Mat mask = Mat::zeros(image.size(), CV_8UC1);
rectangle(mask, Point(window.x, window.y), Point(window.x + window.width,
window.y + window.height), Scalar(255), CV_FILLED);
dd->operator()(prev_image, mask, prev_keypoints, prev_desc);
window.x -= params.window_size;
window.y -= params.window_size;
window.width += params.window_size;
window.height += params.window_size;
rectangle(mask, Point(window.x, window.y), Point(window.x + window.width,
window.y + window.height), Scalar(255), CV_FILLED);
dd->operator()(image, mask, curr_keypoints, curr_desc);
if (prev_keypoints.size() > 4 && curr_keypoints.size() > 4)
{
//descriptor->compute(prev_image, prev_keypoints, prev_desc);
//descriptor->compute(image, curr_keypoints, curr_desc);
matcher->match(prev_desc, curr_desc, matches);
for (int i = 0; i < (int)matches.size(); i++)
{
prev_keys.push_back(prev_keypoints[matches[i].queryIdx].pt);
curr_keys.push_back(curr_keypoints[matches[i].trainIdx].pt);
}
Mat T = findHomography(prev_keys, curr_keys, LMEDS);
prev_trackwindow.x += cvRound(T.at<double> (0, 2));
prev_trackwindow.y += cvRound(T.at<double> (1, 2));
}
prev_center.x = prev_trackwindow.x;
prev_center.y = prev_trackwindow.y;
prev_image = image;
return prev_trackwindow;
}
Rect CvFeatureTracker::updateTrackingWindowWithFlow(Mat image)
{
ittr++;
Size subPixWinSize(10,10), winSize(31,31);
Mat image_bw;
TermCriteria termcrit(TermCriteria::COUNT | TermCriteria::EPS, 20, 0.03);
std::vector<uchar> status;
std::vector<float> err;
cvtColor(image, image_bw, COLOR_BGR2GRAY);
cvtColor(prev_image, prev_image_bw, COLOR_BGR2GRAY);
if (ittr == 1)
{
Mat mask = Mat::zeros(image.size(), CV_8UC1);
rectangle(mask, Point(prev_trackwindow.x, prev_trackwindow.y), Point(
prev_trackwindow.x + prev_trackwindow.width, prev_trackwindow.y
+ prev_trackwindow.height), Scalar(255), CV_FILLED);
goodFeaturesToTrack(image_bw, features[1], 500, 0.01, 20, mask, 3, 0, 0.04);
cornerSubPix(image_bw, features[1], subPixWinSize, Size(-1, -1), termcrit);
}
else
{
calcOpticalFlowPyrLK(prev_image_bw, image_bw, features[0], features[1],
status, err, winSize, 3, termcrit);
Point2f feature0_center(0, 0);
Point2f feature1_center(0, 0);
int goodtracks = 0;
for (int i = 0; i < (int)features[1].size(); i++)
{
if (status[i] == 1)
{
feature0_center.x += features[0][i].x;
feature0_center.y += features[0][i].y;
feature1_center.x += features[1][i].x;
feature1_center.y += features[1][i].y;
goodtracks++;
}
}
feature0_center.x /= goodtracks;
feature0_center.y /= goodtracks;
feature1_center.x /= goodtracks;
feature1_center.y /= goodtracks;
prev_center.x += (feature1_center.x - feature0_center.x);
prev_center.y += (feature1_center.y - feature0_center.y);
prev_trackwindow.x = (int)prev_center.x;
prev_trackwindow.y = (int)prev_center.y;
}
swap(features[0], features[1]);
image.copyTo(prev_image);
return prev_trackwindow;
}
void CvFeatureTracker::setTrackingWindow(Rect _window)
{
prev_trackwindow = _window;
}
Rect CvFeatureTracker::getTrackingWindow()
{
return prev_trackwindow;
}
Point2f CvFeatureTracker::getTrackingCenter()
{
Point2f center(0, 0);
center.x = (float)(prev_center.x + prev_trackwindow.width/2.0);
center.y = (float)(prev_center.y + prev_trackwindow.height/2.0);
return center;
}

721
modules/contrib/src/fuzzymeanshifttracker.cpp
View File

@@ -1,721 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install, copy or use the software.
//
// Copyright (C) 2009, Farhad Dadgostar
// Intel Corporation and third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of Intel Corporation may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "precomp.hpp"
#include "opencv2/contrib/compat.hpp"
CvFuzzyPoint::CvFuzzyPoint(double _x, double _y)
{
x = _x;
y = _y;
}
bool CvFuzzyCurve::between(double x, double x1, double x2)
{
if ((x >= x1) && (x <= x2))
return true;
else if ((x >= x2) && (x <= x1))
return true;
return false;
}
CvFuzzyCurve::CvFuzzyCurve()
{
value = 0;
}
CvFuzzyCurve::~CvFuzzyCurve()
{
// nothing to do
}
void CvFuzzyCurve::setCentre(double _centre)
{
centre = _centre;
}
double CvFuzzyCurve::getCentre()
{
return centre;
}
void CvFuzzyCurve::clear()
{
points.clear();
}
void CvFuzzyCurve::addPoint(double x, double y)
{
points.push_back(CvFuzzyPoint(x, y));
}
double CvFuzzyCurve::calcValue(double param)
{
int size = (int)points.size();
double x1, y1, x2, y2, m, y;
for (int i = 1; i < size; i++)
{
x1 = points[i-1].x;
x2 = points[i].x;
if (between(param, x1, x2)) {
y1 = points[i-1].y;
y2 = points[i].y;
if (x2 == x1)
return y2;
m = (y2-y1)/(x2-x1);
y = m*(param-x1)+y1;
return y;
}
}
return 0;
}
double CvFuzzyCurve::getValue()
{
return value;
}
void CvFuzzyCurve::setValue(double _value)
{
value = _value;
}
CvFuzzyFunction::CvFuzzyFunction()
{
// nothing to do
}
CvFuzzyFunction::~CvFuzzyFunction()
{
curves.clear();
}
void CvFuzzyFunction::addCurve(CvFuzzyCurve *curve, double value)
{
curves.push_back(*curve);
curve->setValue(value);
}
void CvFuzzyFunction::resetValues()
{
int numCurves = (int)curves.size();
for (int i = 0; i < numCurves; i++)
curves[i].setValue(0);
}
double CvFuzzyFunction::calcValue()
{
double s1 = 0, s2 = 0, v;
int numCurves = (int)curves.size();
for (int i = 0; i < numCurves; i++)
{
v = curves[i].getValue();
s1 += curves[i].getCentre() * v;
s2 += v;
}
if (s2 != 0)
return s1/s2;
else
return 0;
}
CvFuzzyCurve *CvFuzzyFunction::newCurve()
{
CvFuzzyCurve *c;
c = new CvFuzzyCurve();
addCurve(c);
return c;
}
CvFuzzyRule::CvFuzzyRule()
{
fuzzyInput1 = NULL;
fuzzyInput2 = NULL;
fuzzyOutput = NULL;
}
CvFuzzyRule::~CvFuzzyRule()
{
if (fuzzyInput1 != NULL)
delete fuzzyInput1;
if (fuzzyInput2 != NULL)
delete fuzzyInput2;
if (fuzzyOutput != NULL)
delete fuzzyOutput;
}
void CvFuzzyRule::setRule(CvFuzzyCurve *c1, CvFuzzyCurve *c2, CvFuzzyCurve *o1)
{
fuzzyInput1 = c1;
fuzzyInput2 = c2;
fuzzyOutput = o1;
}
double CvFuzzyRule::calcValue(double param1, double param2)
{
double v1, v2;
v1 = fuzzyInput1->calcValue(param1);
if (fuzzyInput2 != NULL)
{
v2 = fuzzyInput2->calcValue(param2);
if (v1 < v2)
return v1;
else
return v2;
}
else
return v1;
}
CvFuzzyCurve *CvFuzzyRule::getOutputCurve()
{
return fuzzyOutput;
}
CvFuzzyController::CvFuzzyController()
{
// nothing to do
}
CvFuzzyController::~CvFuzzyController()
{
int size = (int)rules.size();
for(int i = 0; i < size; i++)
delete rules[i];
}
void CvFuzzyController::addRule(CvFuzzyCurve *c1, CvFuzzyCurve *c2, CvFuzzyCurve *o1)
{
CvFuzzyRule *f = new CvFuzzyRule();
rules.push_back(f);
f->setRule(c1, c2, o1);
}
double CvFuzzyController::calcOutput(double param1, double param2)
{
double v;
CvFuzzyFunction list;
int size = (int)rules.size();
for(int i = 0; i < size; i++)
{
v = rules[i]->calcValue(param1, param2);
if (v != 0)
list.addCurve(rules[i]->getOutputCurve(), v);
}
v = list.calcValue();
return v;
}
CvFuzzyMeanShiftTracker::FuzzyResizer::FuzzyResizer()
{
CvFuzzyCurve *i1L, *i1M, *i1H;
CvFuzzyCurve *oS, *oZE, *oE;
CvFuzzyCurve *c;
double MedStart = 0.1, MedWidth = 0.15;
c = iInput.newCurve();
c->addPoint(0, 1);
c->addPoint(0.1, 0);
c->setCentre(0);
i1L = c;
c = iInput.newCurve();
c->addPoint(0.05, 0);
c->addPoint(MedStart, 1);
c->addPoint(MedStart+MedWidth, 1);
c->addPoint(MedStart+MedWidth+0.05, 0);
c->setCentre(MedStart+(MedWidth/2));
i1M = c;
c = iInput.newCurve();
c->addPoint(MedStart+MedWidth, 0);
c->addPoint(1, 1);
c->addPoint(1000, 1);
c->setCentre(1);
i1H = c;
c = iOutput.newCurve();
c->addPoint(-10000, 1);
c->addPoint(-5, 1);
c->addPoint(-0.5, 0);
c->setCentre(-5);
oS = c;
c = iOutput.newCurve();
c->addPoint(-1, 0);
c->addPoint(-0.05, 1);
c->addPoint(0.05, 1);
c->addPoint(1, 0);
c->setCentre(0);
oZE = c;
c = iOutput.newCurve();
c->addPoint(-0.5, 0);
c->addPoint(5, 1);
c->addPoint(1000, 1);
c->setCentre(5);
oE = c;
fuzzyController.addRule(i1L, NULL, oS);
fuzzyController.addRule(i1M, NULL, oZE);
fuzzyController.addRule(i1H, NULL, oE);
}
int CvFuzzyMeanShiftTracker::FuzzyResizer::calcOutput(double edgeDensity, double density)
{
return (int)fuzzyController.calcOutput(edgeDensity, density);
}
CvFuzzyMeanShiftTracker::SearchWindow::SearchWindow()
{
x = 0;
y = 0;
width = 0;
height = 0;
maxWidth = 0;
maxHeight = 0;
xGc = 0;
yGc = 0;
m00 = 0;
m01 = 0;
m10 = 0;
m11 = 0;
m02 = 0;
m20 = 0;
ellipseHeight = 0;
ellipseWidth = 0;
ellipseAngle = 0;
density = 0;
depthLow = 0;
depthHigh = 0;
fuzzyResizer = NULL;
}
CvFuzzyMeanShiftTracker::SearchWindow::~SearchWindow()
{
if (fuzzyResizer != NULL)
delete fuzzyResizer;
}
void CvFuzzyMeanShiftTracker::SearchWindow::setSize(int _x, int _y, int _width, int _height)
{
x = _x;
y = _y;
width = _width;
height = _height;
if (x < 0)
x = 0;
if (y < 0)
y = 0;
if (x + width > maxWidth)
width = maxWidth - x;
if (y + height > maxHeight)
height = maxHeight - y;
}
void CvFuzzyMeanShiftTracker::SearchWindow::initDepthValues(IplImage *maskImage, IplImage *depthMap)
{
unsigned int d=0, mind = 0xFFFF, maxd = 0, m0 = 0, m1 = 0, mc, dd;
unsigned char *data = NULL;
unsigned short *depthData = NULL;
for (int j = 0; j < height; j++)
{
data = (unsigned char *)(maskImage->imageData + (maskImage->widthStep * (j + y)) + x);
if (depthMap)
depthData = (unsigned short *)(depthMap->imageData + (depthMap->widthStep * (j + y)) + x);
for (int i = 0; i < width; i++)
{
if (*data)
{
m0 += 1;
if (depthData)
{
if (*depthData)
{
d = *depthData;
m1 += d;
if (d < mind)
mind = d;
if (d > maxd)
maxd = d;
}
depthData++;
}
}
data++;
}
}
if (m0 > 0)
{
mc = m1/m0;
if ((mc - mind) > (maxd - mc))
dd = maxd - mc;
else
dd = mc - mind;
dd = dd - dd/10;
depthHigh = mc + dd;
depthLow = mc - dd;
}
else
{
depthHigh = 32000;
depthLow = 0;
}
}
bool CvFuzzyMeanShiftTracker::SearchWindow::shift()
{
if ((xGc != (width/2)) || (yGc != (height/2)))
{
setSize(x + (xGc-(width/2)), y + (yGc-(height/2)), width, height);
return true;
}
else
{
return false;
}
}
void CvFuzzyMeanShiftTracker::SearchWindow::extractInfo(IplImage *maskImage, IplImage *depthMap, bool initDepth)
{
m00 = 0;
m10 = 0;
m01 = 0;
m11 = 0;
density = 0;
m02 = 0;
m20 = 0;
ellipseHeight = 0;
ellipseWidth = 0;
maxWidth = maskImage->width;
maxHeight = maskImage->height;
if (initDepth)
initDepthValues(maskImage, depthMap);
unsigned char *maskData = NULL;
unsigned short *depthData = NULL, depth;
bool isOk;
unsigned long count;
verticalEdgeLeft = 0;
verticalEdgeRight = 0;
horizontalEdgeTop = 0;
horizontalEdgeBottom = 0;
for (int j = 0; j < height; j++)
{
maskData = (unsigned char *)(maskImage->imageData + (maskImage->widthStep * (j + y)) + x);
if (depthMap)
depthData = (unsigned short *)(depthMap->imageData + (depthMap->widthStep * (j + y)) + x);
count = 0;
for (int i = 0; i < width; i++)
{
if (*maskData)
{
isOk = true;
if (depthData)
{
depth = (*depthData);
if ((depth > depthHigh) || (depth < depthLow))
isOk = false;
depthData++;
}
if (isOk)
{
m00++;
m01 += j;
m10 += i;
m02 += (j * j);
m20 += (i * i);
m11 += (j * i);
if (i == 0)
verticalEdgeLeft++;
else if (i == width-1)
verticalEdgeRight++;
else if (j == 0)
horizontalEdgeTop++;
else if (j == height-1)
horizontalEdgeBottom++;
count++;
}
}
maskData++;
}
}
if (m00 > 0)
{
xGc = (m10 / m00);
yGc = (m01 / m00);
double a, b, c, e1, e2, e3;
a = ((double)m20/(double)m00)-(xGc * xGc);
b = 2*(((double)m11/(double)m00)-(xGc * yGc));
c = ((double)m02/(double)m00)-(yGc * yGc);
e1 = a+c;
e3 = a-c;
e2 = sqrt((b*b)+(e3*e3));
ellipseHeight = int(sqrt(0.5*(e1+e2)));
ellipseWidth = int(sqrt(0.5*(e1-e2)));
if (e3 == 0)
ellipseAngle = 0;
else
ellipseAngle = 0.5*atan(b/e3);
density = (double)m00/(double)(width * height);
}
else
{
xGc = width / 2;
yGc = height / 2;
ellipseHeight = 0;
ellipseWidth = 0;
ellipseAngle = 0;
density = 0;
}
}
void CvFuzzyMeanShiftTracker::SearchWindow::getResizeAttribsEdgeDensityLinear(int &resizeDx, int &resizeDy, int &resizeDw, int &resizeDh) {
int x1 = horizontalEdgeTop;
int x2 = horizontalEdgeBottom;
int y1 = verticalEdgeLeft;
int y2 = verticalEdgeRight;
int gx = (width*2)/5;
int gy = (height*2)/5;
int lx = width/10;
int ly = height/10;
resizeDy = 0;
resizeDh = 0;
resizeDx = 0;
resizeDw = 0;
if (x1 > gx) {
resizeDy = -1;
} else if (x1 < lx) {
resizeDy = +1;
}
if (x2 > gx) {
resizeDh = resizeDy + 1;
} else if (x2 < lx) {
resizeDh = - (resizeDy + 1);
} else {
resizeDh = - resizeDy;
}
if (y1 > gy) {
resizeDx = -1;
} else if (y1 < ly) {
resizeDx = +1;
}
if (y2 > gy) {
resizeDw = resizeDx + 1;
} else if (y2 < ly) {
resizeDw = - (resizeDx + 1);
} else {
resizeDw = - resizeDx;
}
}
void CvFuzzyMeanShiftTracker::SearchWindow::getResizeAttribsInnerDensity(int &resizeDx, int &resizeDy, int &resizeDw, int &resizeDh)
{
int newWidth, newHeight, dx, dy;
double px, py;
newWidth = int(sqrt(double(m00)*1.3));
newHeight = int(newWidth*1.2);
dx = (newWidth - width);
dy = (newHeight - height);
px = (double)xGc/(double)width;
py = (double)yGc/(double)height;
resizeDx = (int)(px*dx);
resizeDy = (int)(py*dy);
resizeDw = (int)((1-px)*dx);
resizeDh = (int)((1-py)*dy);
}
void CvFuzzyMeanShiftTracker::SearchWindow::getResizeAttribsEdgeDensityFuzzy(int &resizeDx, int &resizeDy, int &resizeDw, int &resizeDh)
{
double dx1=0, dx2, dy1, dy2;
resizeDy = 0;
resizeDh = 0;
resizeDx = 0;
resizeDw = 0;
if (fuzzyResizer == NULL)
fuzzyResizer = new FuzzyResizer();
dx2 = fuzzyResizer->calcOutput(double(verticalEdgeRight)/double(height), density);
if (dx1 == dx2)
{
resizeDx = int(-dx1);
resizeDw = int(dx1+dx2);
}
dy1 = fuzzyResizer->calcOutput(double(horizontalEdgeTop)/double(width), density);
dy2 = fuzzyResizer->calcOutput(double(horizontalEdgeBottom)/double(width), density);
dx1 = fuzzyResizer->calcOutput(double(verticalEdgeLeft)/double(height), density);
dx2 = fuzzyResizer->calcOutput(double(verticalEdgeRight)/double(height), density);
//if (dx1 == dx2)
{
resizeDx = int(-dx1);
resizeDw = int(dx1+dx2);
}
dy1 = fuzzyResizer->calcOutput(double(horizontalEdgeTop)/double(width), density);
dy2 = fuzzyResizer->calcOutput(double(horizontalEdgeBottom)/double(width), density);
//if (dy1 == dy2)
{
resizeDy = int(-dy1);
resizeDh = int(dy1+dy2);
}
}
bool CvFuzzyMeanShiftTracker::SearchWindow::meanShift(IplImage *maskImage, IplImage *depthMap, int maxIteration, bool initDepth)
{
numShifts = 0;
do
{
extractInfo(maskImage, depthMap, initDepth);
if (! shift())
return true;
} while (++numShifts < maxIteration);
return false;
}
void CvFuzzyMeanShiftTracker::findOptimumSearchWindow(SearchWindow &searchWindow, IplImage *maskImage, IplImage *depthMap, int maxIteration, int resizeMethod, bool initDepth)
{
int resizeDx, resizeDy, resizeDw, resizeDh;
resizeDx = 0;
resizeDy = 0;
resizeDw = 0;
resizeDh = 0;
searchWindow.numIters = 0;
for (int i = 0; i < maxIteration; i++)
{
searchWindow.numIters++;
searchWindow.meanShift(maskImage, depthMap, MaxMeanShiftIteration, initDepth);
switch (resizeMethod)
{
case rmEdgeDensityLinear :
searchWindow.getResizeAttribsEdgeDensityLinear(resizeDx, resizeDy, resizeDw, resizeDh);
break;
case rmEdgeDensityFuzzy :
//searchWindow.getResizeAttribsEdgeDensityLinear(resizeDx, resizeDy, resizeDw, resizeDh);
searchWindow.getResizeAttribsEdgeDensityFuzzy(resizeDx, resizeDy, resizeDw, resizeDh);
break;
case rmInnerDensity :
searchWindow.getResizeAttribsInnerDensity(resizeDx, resizeDy, resizeDw, resizeDh);
break;
default:
searchWindow.getResizeAttribsEdgeDensityLinear(resizeDx, resizeDy, resizeDw, resizeDh);
}
searchWindow.ldx = resizeDx;
searchWindow.ldy = resizeDy;
searchWindow.ldw = resizeDw;
searchWindow.ldh = resizeDh;
if ((resizeDx == 0) && (resizeDy == 0) && (resizeDw == 0) && (resizeDh == 0))
break;
searchWindow.setSize(searchWindow.x + resizeDx, searchWindow.y + resizeDy, searchWindow.width + resizeDw, searchWindow.height + resizeDh);
}
}
CvFuzzyMeanShiftTracker::CvFuzzyMeanShiftTracker()
{
searchMode = tsSetWindow;
}
CvFuzzyMeanShiftTracker::~CvFuzzyMeanShiftTracker()
{
// nothing to do
}
void CvFuzzyMeanShiftTracker::track(IplImage *maskImage, IplImage *depthMap, int resizeMethod, bool resetSearch, int minKernelMass)
{
bool initDepth = false;
if (resetSearch)
searchMode = tsSetWindow;
switch (searchMode)
{
case tsDisabled:
return;
case tsSearching:
return;
case tsSetWindow:
kernel.maxWidth = maskImage->width;
kernel.maxHeight = maskImage->height;
kernel.setSize(0, 0, maskImage->width, maskImage->height);
initDepth = true;
case tsTracking:
searchMode = tsSearching;
findOptimumSearchWindow(kernel, maskImage, depthMap, MaxSetSizeIteration, resizeMethod, initDepth);
if ((kernel.density == 0) || (kernel.m00 < minKernelMass))
searchMode = tsSetWindow;
else
searchMode = tsTracking;
}
}

139
modules/contrib/src/gencolors.cpp
View File

@@ -1,139 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "precomp.hpp"
#include <iostream>
using namespace cv;
static void downsamplePoints( const Mat& src, Mat& dst, size_t count )
{
CV_Assert( count >= 2 );
CV_Assert( src.cols == 1 || src.rows == 1 );
CV_Assert( src.total() >= count );
CV_Assert( src.type() == CV_8UC3);
dst.create( 1, (int)count, CV_8UC3 );
//TODO: optimize by exploiting symmetry in the distance matrix
Mat dists( (int)src.total(), (int)src.total(), CV_32FC1, Scalar(0) );
if( dists.empty() )
std::cerr << "Such big matrix cann't be created." << std::endl;
for( int i = 0; i < dists.rows; i++ )
{
for( int j = i; j < dists.cols; j++ )
{
float dist = (float)norm(src.at<Point3_<uchar> >(i) - src.at<Point3_<uchar> >(j));
dists.at<float>(j, i) = dists.at<float>(i, j) = dist;
}
}
double maxVal;
Point maxLoc;
minMaxLoc(dists, 0, &maxVal, 0, &maxLoc);
dst.at<Point3_<uchar> >(0) = src.at<Point3_<uchar> >(maxLoc.x);
dst.at<Point3_<uchar> >(1) = src.at<Point3_<uchar> >(maxLoc.y);
Mat activedDists( 0, dists.cols, dists.type() );
Mat candidatePointsMask( 1, dists.cols, CV_8UC1, Scalar(255) );
activedDists.push_back( dists.row(maxLoc.y) );
candidatePointsMask.at<uchar>(0, maxLoc.y) = 0;
for( size_t i = 2; i < count; i++ )
{
activedDists.push_back(dists.row(maxLoc.x));
candidatePointsMask.at<uchar>(0, maxLoc.x) = 0;
Mat minDists;
reduce( activedDists, minDists, 0, REDUCE_MIN );
minMaxLoc( minDists, 0, &maxVal, 0, &maxLoc, candidatePointsMask );
dst.at<Point3_<uchar> >((int)i) = src.at<Point3_<uchar> >(maxLoc.x);
}
}
void cv::generateColors( std::vector<Scalar>& colors, size_t count, size_t factor )
{
if( count < 1 )
return;
colors.resize(count);
if( count == 1 )
{
colors[0] = Scalar(0,0,255); // red
return;
}
if( count == 2 )
{
colors[0] = Scalar(0,0,255); // red
colors[1] = Scalar(0,255,0); // green
return;
}
// Generate a set of colors in RGB space. A size of the set is severel times (=factor) larger then
// the needed count of colors.
Mat bgr( 1, (int)(count*factor), CV_8UC3 );
randu( bgr, 0, 256 );
// Convert the colors set to Lab space.
// Distances between colors in this space correspond a human perception.
Mat lab;
cvtColor( bgr, lab, COLOR_BGR2Lab);
// Subsample colors from the generated set so that
// to maximize the minimum distances between each other.
// Douglas-Peucker algorithm is used for this.
Mat lab_subset;
downsamplePoints( lab, lab_subset, count );
// Convert subsampled colors back to RGB
Mat bgr_subset;
cvtColor( lab_subset, bgr_subset, COLOR_Lab2BGR );
CV_Assert( bgr_subset.total() == count );
for( size_t i = 0; i < count; i++ )
{
Point3_<uchar> c = bgr_subset.at<Point3_<uchar> >((int)i);
colors[i] = Scalar(c.x, c.y, c.z);
}
}

235
modules/contrib/src/hybridtracker.cpp
View File

@@ -1,235 +0,0 @@
//*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2008-2011, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of Intel Corporation may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "precomp.hpp"
#include "opencv2/contrib/hybridtracker.hpp"
using namespace cv;
CvHybridTrackerParams::CvHybridTrackerParams(float _ft_tracker_weight, float _ms_tracker_weight,
CvFeatureTrackerParams _ft_params,
CvMeanShiftTrackerParams _ms_params,
CvMotionModel)
{
ft_tracker_weight = _ft_tracker_weight;
ms_tracker_weight = _ms_tracker_weight;
ft_params = _ft_params;
ms_params = _ms_params;
}
CvMeanShiftTrackerParams::CvMeanShiftTrackerParams(int _tracking_type, CvTermCriteria _term_crit)
{
tracking_type = _tracking_type;
term_crit = _term_crit;
}
CvHybridTracker::CvHybridTracker() {
}
CvHybridTracker::CvHybridTracker(HybridTrackerParams _params) :
params(_params) {
params.ft_params.feature_type = CvFeatureTrackerParams::SIFT;
mstracker = new CvMeanShiftTracker(params.ms_params);
fttracker = new CvFeatureTracker(params.ft_params);
}
CvHybridTracker::~CvHybridTracker() {
if (mstracker != NULL)
delete mstracker;
if (fttracker != NULL)
delete fttracker;
}
inline float CvHybridTracker::getL2Norm(Point2f p1, Point2f p2) {
float distance = (p1.x - p2.x) * (p1.x - p2.x) + (p1.y - p2.y) * (p1.y
- p2.y);
return std::sqrt(distance);
}
Mat CvHybridTracker::getDistanceProjection(Mat image, Point2f center) {
Mat hist(image.size(), CV_64F);
double lu = getL2Norm(Point(0, 0), center);
double ru = getL2Norm(Point(0, image.size().width), center);
double rd = getL2Norm(Point(image.size().height, image.size().width),
center);
double ld = getL2Norm(Point(image.size().height, 0), center);
double max = (lu < ru) ? lu : ru;
max = (max < rd) ? max : rd;
max = (max < ld) ? max : ld;
for (int i = 0; i < hist.rows; i++)
for (int j = 0; j < hist.cols; j++)
hist.at<double> (i, j) = 1.0 - (getL2Norm(Point(i, j), center)
/ max);
return hist;
}
Mat CvHybridTracker::getGaussianProjection(Mat image, int ksize, double sigma,
Point2f center) {
Mat kernel = getGaussianKernel(ksize, sigma, CV_64F);
double max = kernel.at<double> (ksize / 2);
Mat hist(image.size(), CV_64F);
for (int i = 0; i < hist.rows; i++)
for (int j = 0; j < hist.cols; j++) {
int pos = cvRound(getL2Norm(Point(i, j), center));
if (pos < ksize / 2.0)
hist.at<double> (i, j) = 1.0 - (kernel.at<double> (pos) / max);
}
return hist;
}
void CvHybridTracker::newTracker(Mat image, Rect selection) {
prev_proj = Mat::zeros(image.size(), CV_64FC1);
prev_center = Point2f(selection.x + selection.width / 2.0f, selection.y
+ selection.height / 2.0f);
prev_window = selection;
mstracker->newTrackingWindow(image, selection);
fttracker->newTrackingWindow(image, selection);
samples = cvCreateMat(2, 1, CV_32FC1);
labels = cvCreateMat(2, 1, CV_32SC1);
ittr = 0;
}
void CvHybridTracker::updateTracker(Mat image) {
ittr++;
//copy over clean images: TODO
mstracker->updateTrackingWindow(image);
fttracker->updateTrackingWindowWithFlow(image);
if (params.motion_model == CvMotionModel::EM)
updateTrackerWithEM(image);
else
updateTrackerWithLowPassFilter(image);
// Regression to find new weights
Point2f ms_center = mstracker->getTrackingEllipse().center;
Point2f ft_center = fttracker->getTrackingCenter();
#ifdef DEBUG_HYTRACKER
circle(image, ms_center, 3, Scalar(0, 0, 255), -1, 8);
circle(image, ft_center, 3, Scalar(255, 0, 0), -1, 8);
putText(image, "ms", Point(ms_center.x+2, ms_center.y), FONT_HERSHEY_PLAIN, 0.75, Scalar(255, 255, 255));
putText(image, "ft", Point(ft_center.x+2, ft_center.y), FONT_HERSHEY_PLAIN, 0.75, Scalar(255, 255, 255));
#endif
double ms_len = getL2Norm(ms_center, curr_center);
double ft_len = getL2Norm(ft_center, curr_center);
double total_len = ms_len + ft_len;
params.ms_tracker_weight *= (ittr - 1);
params.ms_tracker_weight += (float)((ms_len / total_len));
params.ms_tracker_weight /= ittr;
params.ft_tracker_weight *= (ittr - 1);
params.ft_tracker_weight += (float)((ft_len / total_len));
params.ft_tracker_weight /= ittr;
circle(image, prev_center, 3, Scalar(0, 0, 0), -1, 8);
circle(image, curr_center, 3, Scalar(255, 255, 255), -1, 8);
prev_center = curr_center;
prev_window.x = (int)(curr_center.x-prev_window.width/2.0);
prev_window.y = (int)(curr_center.y-prev_window.height/2.0);
mstracker->setTrackingWindow(prev_window);
fttracker->setTrackingWindow(prev_window);
}
void CvHybridTracker::updateTrackerWithEM(Mat image) {
Mat ms_backproj = mstracker->getHistogramProjection(CV_64F);
Mat ms_distproj = getDistanceProjection(image, mstracker->getTrackingCenter());
Mat ms_proj = ms_backproj.mul(ms_distproj);
float dist_err = getL2Norm(mstracker->getTrackingCenter(), fttracker->getTrackingCenter());
Mat ft_gaussproj = getGaussianProjection(image, cvRound(dist_err), -1, fttracker->getTrackingCenter());
Mat ft_distproj = getDistanceProjection(image, fttracker->getTrackingCenter());
Mat ft_proj = ft_gaussproj.mul(ft_distproj);
Mat proj = params.ms_tracker_weight * ms_proj + params.ft_tracker_weight * ft_proj + prev_proj;
int sample_count = countNonZero(proj);
cvReleaseMat(&samples);
cvReleaseMat(&labels);
samples = cvCreateMat(sample_count, 2, CV_32FC1);
labels = cvCreateMat(sample_count, 1, CV_32SC1);
int count = 0;
for (int i = 0; i < proj.rows; i++)
for (int j = 0; j < proj.cols; j++)
if (proj.at<double> (i, j) > 0) {
samples->data.fl[count * 2] = (float)i;
samples->data.fl[count * 2 + 1] = (float)j;
count++;
}
cv::Mat lbls;
EM em_model(1, EM::COV_MAT_SPHERICAL, TermCriteria(TermCriteria::COUNT + TermCriteria::EPS, 10000, 0.001));
em_model.train(cvarrToMat(samples), noArray(), lbls);
if(labels)
lbls.copyTo(cvarrToMat(labels));
Mat em_means = em_model.get<Mat>("means");
curr_center.x = (float)em_means.at<float>(0, 0);
curr_center.y = (float)em_means.at<float>(0, 1);
}
void CvHybridTracker::updateTrackerWithLowPassFilter(Mat) {
RotatedRect ms_track = mstracker->getTrackingEllipse();
Point2f ft_center = fttracker->getTrackingCenter();
float a = params.low_pass_gain;
curr_center.x = (1 - a) * prev_center.x + a * (params.ms_tracker_weight * ms_track.center.x + params.ft_tracker_weight * ft_center.x);
curr_center.y = (1 - a) * prev_center.y + a * (params.ms_tracker_weight * ms_track.center.y + params.ft_tracker_weight * ft_center.y);
}
Rect CvHybridTracker::getTrackingWindow() {
return prev_window;
}

204
modules/contrib/src/inputoutput.cpp
View File

@@ -1,204 +0,0 @@
#include "opencv2/contrib.hpp"
#include "cvconfig.h"
#if defined(WIN32) || defined(_WIN32)
#include <windows.h>
#include <tchar.h>
#else
#include <dirent.h>
#endif
namespace cv
{
std::vector<String> Directory::GetListFiles( const String& path, const String & exten, bool addPath )
{
std::vector<String> list;
list.clear();
String path_f = path + "/" + exten;
#ifdef WIN32
#ifdef HAVE_WINRT
WIN32_FIND_DATAW FindFileData;
#else
WIN32_FIND_DATAA FindFileData;
#endif
HANDLE hFind;
#ifdef HAVE_WINRT
wchar_t wpath[MAX_PATH];
size_t copied = mbstowcs(wpath, path_f.c_str(), MAX_PATH);
CV_Assert((copied != MAX_PATH) && (copied != (size_t)-1));
hFind = FindFirstFileExW(wpath, FindExInfoStandard, &FindFileData, FindExSearchNameMatch, NULL, 0);
#else
hFind = FindFirstFileA((LPCSTR)path_f.c_str(), &FindFileData);
#endif
if (hFind == INVALID_HANDLE_VALUE)
{
return list;
}
else
{
do
{
if (FindFileData.dwFileAttributes == FILE_ATTRIBUTE_NORMAL ||
FindFileData.dwFileAttributes == FILE_ATTRIBUTE_ARCHIVE ||
FindFileData.dwFileAttributes == FILE_ATTRIBUTE_HIDDEN ||
FindFileData.dwFileAttributes == FILE_ATTRIBUTE_SYSTEM ||
FindFileData.dwFileAttributes == FILE_ATTRIBUTE_READONLY)
{
char* fname;
#ifdef HAVE_WINRT
char fname_tmp[MAX_PATH] = {0};
size_t copied = wcstombs(fname_tmp, FindFileData.cFileName, MAX_PATH);
CV_Assert((copied != MAX_PATH) && (copied != (size_t)-1));
fname = fname_tmp;
#else
fname = FindFileData.cFileName;
#endif
if (addPath)
list.push_back(path + "/" + String(fname));
else
list.push_back(String(fname));
}
}
#ifdef HAVE_WINRT
while(FindNextFileW(hFind, &FindFileData));
#else
while(FindNextFileA(hFind, &FindFileData));
#endif
FindClose(hFind);
}
#else
(void)addPath;
DIR *dp;
struct dirent *dirp;
if((dp = opendir(path.c_str())) == NULL)
{
return list;
}
while ((dirp = readdir(dp)) != NULL)
{
if (dirp->d_type == DT_REG)
{
if (exten.compare("*") == 0)
list.push_back(static_cast<String>(dirp->d_name));
else
if (String(dirp->d_name).find(exten) != String::npos)
list.push_back(static_cast<String>(dirp->d_name));
}
}
closedir(dp);
#endif
return list;
}
std::vector<String> Directory::GetListFolders( const String& path, const String & exten, bool addPath )
{
std::vector<String> list;
String path_f = path + "/" + exten;
list.clear();
#ifdef WIN32
#ifdef HAVE_WINRT
WIN32_FIND_DATAW FindFileData;
#else
WIN32_FIND_DATAA FindFileData;
#endif
HANDLE hFind;
#ifdef HAVE_WINRT
wchar_t wpath [MAX_PATH];
size_t copied = mbstowcs(wpath, path_f.c_str(), path_f.size());
CV_Assert((copied != MAX_PATH) && (copied != (size_t)-1));
hFind = FindFirstFileExW(wpath, FindExInfoStandard, &FindFileData, FindExSearchNameMatch, NULL, 0);
#else
hFind = FindFirstFileA((LPCSTR)path_f.c_str(), &FindFileData);
#endif
if (hFind == INVALID_HANDLE_VALUE)
{
return list;
}
else
{
do
{
#ifdef HAVE_WINRT
if (FindFileData.dwFileAttributes == FILE_ATTRIBUTE_DIRECTORY &&
wcscmp(FindFileData.cFileName, L".") != 0 &&
wcscmp(FindFileData.cFileName, L"..") != 0)
#else
if (FindFileData.dwFileAttributes == FILE_ATTRIBUTE_DIRECTORY &&
strcmp(FindFileData.cFileName, ".") != 0 &&
strcmp(FindFileData.cFileName, "..") != 0)
#endif
{
char* fname;
#ifdef HAVE_WINRT
char fname_tmp[MAX_PATH];
size_t copied = wcstombs(fname_tmp, FindFileData.cFileName, MAX_PATH);
CV_Assert((copied != MAX_PATH) && (copied != (size_t)-1));
fname = fname_tmp;
#else
fname = FindFileData.cFileName;
#endif
if (addPath)
list.push_back(path + "/" + String(fname));
else
list.push_back(String(fname));
}
}
#ifdef HAVE_WINRT
while(FindNextFileW(hFind, &FindFileData));
#else
while(FindNextFileA(hFind, &FindFileData));
#endif
FindClose(hFind);
}
#else
(void)addPath;
DIR *dp;
struct dirent *dirp;
if((dp = opendir(path_f.c_str())) == NULL)
{
return list;
}
while ((dirp = readdir(dp)) != NULL)
{
if (dirp->d_type == DT_DIR &&
strcmp(dirp->d_name, ".") != 0 &&
strcmp(dirp->d_name, "..") != 0 )
{
if (exten.compare("*") == 0)
list.push_back(static_cast<String>(dirp->d_name));
else
if (String(dirp->d_name).find(exten) != String::npos)
list.push_back(static_cast<String>(dirp->d_name));
}
}
closedir(dp);
#endif
return list;
}
std::vector<String> Directory::GetListFilesR ( const String& path, const String & exten, bool addPath )
{
std::vector<String> list = Directory::GetListFiles(path, exten, addPath);
std::vector<String> dirs = Directory::GetListFolders(path, exten, addPath);
std::vector<String>::const_iterator it;
for (it = dirs.begin(); it != dirs.end(); ++it)
{
std::vector<String> cl = Directory::GetListFiles(*it, exten, addPath);
list.insert(list.end(), cl.begin(), cl.end());
}
return list;
}
}

1108
modules/contrib/src/lda.cpp
View File

File diff suppressed because it is too large Load Diff

651
modules/contrib/src/logpolar_bsm.cpp
View File

@@ -1,651 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2012, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The names of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
/*******************************************************************************************
The LogPolar Blind Spot Model code has been contributed by Fabio Solari and Manuela Chessa.
More details can be found in:
M. Chessa, S. P. Sabatini, F. Solari and F. Tatti (2011)
A Quantitative Comparison of Speed and Reliability for Log-Polar Mapping Techniques,
Computer Vision Systems - 8th International Conference,
ICVS 2011, Sophia Antipolis, France, September 20-22, 2011
(http://dx.doi.org/10.1007/978-3-642-23968-7_5)
********************************************************************************************/
#include "precomp.hpp"
#include <cmath>
#include <vector>
namespace cv
{
//------------------------------------interp-------------------------------------------
LogPolar_Interp::LogPolar_Interp(int w, int h, Point2i center, int _R, double _ro0, int _interp, int full, int _s, int sp)
{
if ( (center.x!=w/2 || center.y!=h/2) && full==0) full=1;
if (center.x<0) center.x=0;
if (center.y<0) center.y=0;
if (center.x>=w) center.x=w-1;
if (center.y>=h) center.y=h-1;
if (full){
int rtmp;
if (center.x<=w/2 && center.y>=h/2)
rtmp=(int)std::sqrt((float)center.y*center.y + (float)(w-center.x)*(w-center.x));
else if (center.x>=w/2 && center.y>=h/2)
rtmp=(int)std::sqrt((float)center.y*center.y + (float)center.x*center.x);
else if (center.x>=w/2 && center.y<=h/2)
rtmp=(int)std::sqrt((float)(h-center.y)*(h-center.y) + (float)center.x*center.x);
else //if (center.x<=w/2 && center.y<=h/2)
rtmp=(int)std::sqrt((float)(h-center.y)*(h-center.y) + (float)(w-center.x)*(w-center.x));
M=2*rtmp; N=2*rtmp;
top = M/2 - center.y;
bottom = M/2 - (h-center.y);
left = M/2 - center.x;
right = M/2 - (w - center.x);
}else{
top=bottom=left=right=0;
M=w; N=h;
}
if (sp){
int jc=M/2-1, ic=N/2-1;
int _romax=std::min(ic, jc);
double _a=std::exp(std::log((double)(_romax/2-1)/(double)ro0)/(double)R);
S=(int) floor(2*CV_PI/(_a-1)+0.5);
}
interp=_interp;
create_map(M, N, _R, _s, _ro0);
}
void LogPolar_Interp::create_map(int _M, int _n, int _R, int _s, double _ro0)
{
M=_M;
N=_n;
R=_R;
S=_s;
ro0=_ro0;
int jc=N/2-1, ic=M/2-1;
romax=std::min(ic, jc);
a=std::exp(std::log((double)romax/(double)ro0)/(double)R);
q=((double)S)/(2*CV_PI);
Rsri = Mat::zeros(S,R,CV_32FC1);
Csri = Mat::zeros(S,R,CV_32FC1);
ETAyx = Mat::zeros(N,M,CV_32FC1);
CSIyx = Mat::zeros(N,M,CV_32FC1);
for(int v=0; v<S; v++)
{
for(int u=0; u<R; u++)
{
Rsri.at<float>(v,u)=(float)(ro0*std::pow(a,u)*sin(v/q)+jc);
Csri.at<float>(v,u)=(float)(ro0*std::pow(a,u)*cos(v/q)+ic);
}
}
for(int j=0; j<N; j++)
{
for(int i=0; i<M; i++)
{
double theta;
if(i>=ic)
theta=atan((double)(j-jc)/(double)(i-ic));
else
theta=atan((double)(j-jc)/(double)(i-ic))+CV_PI;
if(theta<0)
theta+=2*CV_PI;
ETAyx.at<float>(j,i)=(float)(q*theta);
double ro2=(j-jc)*(j-jc)+(i-ic)*(i-ic);
CSIyx.at<float>(j,i)=(float)(0.5*std::log(ro2/(ro0*ro0))/std::log(a));
}
}
}
const Mat LogPolar_Interp::to_cortical(const Mat &source)
{
Mat out(S,R,CV_8UC1,Scalar(0));
Mat source_border;
copyMakeBorder(source,source_border,top,bottom,left,right,BORDER_CONSTANT,Scalar(0));
remap(source_border,out,Csri,Rsri,interp);
return out;
}
const Mat LogPolar_Interp::to_cartesian(const Mat &source)
{
Mat out(N,M,CV_8UC1,Scalar(0));
Mat source_border;
if (interp==INTER_NEAREST || interp==INTER_LINEAR){
copyMakeBorder(source,source_border,0,1,0,0,BORDER_CONSTANT,Scalar(0));
Mat rowS0 = source_border.row(S);
source_border.row(0).copyTo(rowS0);
} else if (interp==INTER_CUBIC){
copyMakeBorder(source,source_border,0,2,0,0,BORDER_CONSTANT,Scalar(0));
Mat rowS0 = source_border.row(S);
Mat rowS1 = source_border.row(S+1);
source_border.row(0).copyTo(rowS0);
source_border.row(1).copyTo(rowS1);
} else if (interp==INTER_LANCZOS4){
copyMakeBorder(source,source_border,0,4,0,0,BORDER_CONSTANT,Scalar(0));
Mat rowS0 = source_border.row(S);
Mat rowS1 = source_border.row(S+1);
Mat rowS2 = source_border.row(S+2);
Mat rowS3 = source_border.row(S+3);
source_border.row(0).copyTo(rowS0);
source_border.row(1).copyTo(rowS1);
source_border.row(2).copyTo(rowS2);
source_border.row(3).copyTo(rowS3);
}
remap(source_border,out,CSIyx,ETAyx,interp);
Mat out_cropped=out(Range(top,N-1-bottom),Range(left,M-1-right));
return out_cropped;
}
LogPolar_Interp::~LogPolar_Interp()
{
}
//------------------------------------overlapping----------------------------------
LogPolar_Overlapping::LogPolar_Overlapping(int w, int h, Point2i center, int _R, double _ro0, int full, int _s, int sp)
{
if ( (center.x!=w/2 || center.y!=h/2) && full==0) full=1;
if (center.x<0) center.x=0;
if (center.y<0) center.y=0;
if (center.x>=w) center.x=w-1;
if (center.y>=h) center.y=h-1;
if (full){
int rtmp;
if (center.x<=w/2 && center.y>=h/2)
rtmp=(int)std::sqrt((float)center.y*center.y + (float)(w-center.x)*(w-center.x));
else if (center.x>=w/2 && center.y>=h/2)
rtmp=(int)std::sqrt((float)center.y*center.y + (float)center.x*center.x);
else if (center.x>=w/2 && center.y<=h/2)
rtmp=(int)std::sqrt((float)(h-center.y)*(h-center.y) + (float)center.x*center.x);
else //if (center.x<=w/2 && center.y<=h/2)
rtmp=(int)std::sqrt((float)(h-center.y)*(h-center.y) + (float)(w-center.x)*(w-center.x));
M=2*rtmp; N=2*rtmp;
top = M/2 - center.y;
bottom = M/2 - (h-center.y);
left = M/2 - center.x;
right = M/2 - (w - center.x);
}else{
top=bottom=left=right=0;
M=w; N=h;
}
if (sp){
int jc=M/2-1, ic=N/2-1;
int _romax=std::min(ic, jc);
double _a=std::exp(std::log((double)(_romax/2-1)/(double)ro0)/(double)R);
S=(int) floor(2*CV_PI/(_a-1)+0.5);
}
create_map(M, N, _R, _s, _ro0);
}
void LogPolar_Overlapping::create_map(int _M, int _n, int _R, int _s, double _ro0)
{
M=_M;
N=_n;
R=_R;
S=_s;
ro0=_ro0;
int jc=N/2-1, ic=M/2-1;
romax=std::min(ic, jc);
a=std::exp(std::log((double)romax/(double)ro0)/(double)R);
q=((double)S)/(2*CV_PI);
ind1=0;
Rsri=Mat::zeros(S,R,CV_32FC1);
Csri=Mat::zeros(S,R,CV_32FC1);
ETAyx=Mat::zeros(N,M,CV_32FC1);
CSIyx=Mat::zeros(N,M,CV_32FC1);
Rsr.resize(R*S);
Csr.resize(R*S);
Wsr.resize(R);
w_ker_2D.resize(R*S);
for(int v=0; v<S; v++)
{
for(int u=0; u<R; u++)
{
Rsri.at<float>(v,u)=(float)(ro0*std::pow(a,u)*sin(v/q)+jc);
Csri.at<float>(v,u)=(float)(ro0*std::pow(a,u)*cos(v/q)+ic);
Rsr[v*R+u]=(int)floor(Rsri.at<float>(v,u));
Csr[v*R+u]=(int)floor(Csri.at<float>(v,u));
}
}
bool done=false;
for(int i=0; i<R; i++)
{
Wsr[i]=ro0*(a-1)*std::pow(a,i-1);
if((Wsr[i]>1)&&(done==false))
{
ind1=i;
done =true;
}
}
for(int j=0; j<N; j++)
{
for(int i=0; i<M; i++)//mdf
{
double theta;
if(i>=ic)
theta=atan((double)(j-jc)/(double)(i-ic));
else
theta=atan((double)(j-jc)/(double)(i-ic))+CV_PI;
if(theta<0)
theta+=2*CV_PI;
ETAyx.at<float>(j,i)=(float)(q*theta);
double ro2=(j-jc)*(j-jc)+(i-ic)*(i-ic);
CSIyx.at<float>(j,i)=(float)(0.5*std::log(ro2/(ro0*ro0))/std::log(a));
}
}
for(int v=0; v<S; v++)
for(int u=ind1; u<R; u++)
{
//double sigma=Wsr[u]/2.0;
double sigma=Wsr[u]/3.0;//modf
int w=(int) floor(3*sigma+0.5);
w_ker_2D[v*R+u].w=w;
w_ker_2D[v*R+u].weights.resize((2*w+1)*(2*w+1));
double dx=Csri.at<float>(v,u)-Csr[v*R+u];
double dy=Rsri.at<float>(v,u)-Rsr[v*R+u];
double tot=0;
for(int j=0; j<2*w+1; j++)
for(int i=0; i<2*w+1; i++)
{
(w_ker_2D[v*R+u].weights)[j*(2*w+1)+i]=std::exp(-(std::pow(i-w-dx, 2)+std::pow(j-w-dy, 2))/(2*sigma*sigma));
tot+=(w_ker_2D[v*R+u].weights)[j*(2*w+1)+i];
}
for(int j=0; j<(2*w+1); j++)
for(int i=0; i<(2*w+1); i++)
(w_ker_2D[v*R+u].weights)[j*(2*w+1)+i]/=tot;
}
}
const Mat LogPolar_Overlapping::to_cortical(const Mat &source)
{
Mat out(S,R,CV_8UC1,Scalar(0));
Mat source_border;
copyMakeBorder(source,source_border,top,bottom,left,right,BORDER_CONSTANT,Scalar(0));
remap(source_border,out,Csri,Rsri,INTER_LINEAR);
int wm=w_ker_2D[R-1].w;
std::vector<int> IMG((M+2*wm+1)*(N+2*wm+1), 0);
for(int j=0; j<N; j++)
for(int i=0; i<M; i++)
IMG[(M+2*wm+1)*(j+wm)+i+wm]=source_border.at<uchar>(j,i);
for(int v=0; v<S; v++)
for(int u=ind1; u<R; u++)
{
int w=w_ker_2D[v*R+u].w;
double tmp=0;
for(int rf=0; rf<(2*w+1); rf++)
{
for(int cf=0; cf<(2*w+1); cf++)
{
double weight=(w_ker_2D[v*R+u]).weights[rf*(2*w+1)+cf];
tmp+=IMG[(M+2*wm+1)*((rf-w)+Rsr[v*R+u]+wm)+((cf-w)+Csr[v*R+u]+wm)]*weight;
}
}
out.at<uchar>(v,u)=(uchar) floor(tmp+0.5);
}
return out;
}
const Mat LogPolar_Overlapping::to_cartesian(const Mat &source)
{
Mat out(N,M,CV_8UC1,Scalar(0));
Mat source_border;
copyMakeBorder(source,source_border,0,1,0,0,BORDER_CONSTANT,Scalar(0));
Mat rowS = source_border.row(S);
source_border.row(0).copyTo(rowS);
remap(source_border,out,CSIyx,ETAyx,INTER_LINEAR);
int wm=w_ker_2D[R-1].w;
std::vector<double> IMG((N+2*wm+1)*(M+2*wm+1), 0.);
std::vector<double> NOR((N+2*wm+1)*(M+2*wm+1), 0.);
for(int v=0; v<S; v++)
for(int u=ind1; u<R; u++)
{
int w=w_ker_2D[v*R+u].w;
for(int j=0; j<(2*w+1); j++)
{
for(int i=0; i<(2*w+1); i++)
{
int ind=(M+2*wm+1)*((j-w)+Rsr[v*R+u]+wm)+(i-w)+Csr[v*R+u]+wm;
IMG[ind]+=((w_ker_2D[v*R+u]).weights[j*(2*w+1)+i])*source.at<uchar>(v, u);
NOR[ind]+=((w_ker_2D[v*R+u]).weights[j*(2*w+1)+i]);
}
}
}
for(int i=0; i<((N+2*wm+1)*(M+2*wm+1)); i++)
IMG[i]/=NOR[i];
//int xc=M/2-1, yc=N/2-1;
for(int j=wm; j<N+wm; j++)
for(int i=wm; i<M+wm; i++)
{
/*if(NOR[(M+2*wm+1)*j+i]>0)
ret[M*(j-wm)+i-wm]=(int) floor(IMG[(M+2*wm+1)*j+i]+0.5);*/
//int ro=(int)floor(std::sqrt((double)((j-wm-yc)*(j-wm-yc)+(i-wm-xc)*(i-wm-xc))));
int csi=(int) floor(CSIyx.at<float>(j-wm,i-wm));
if((csi>=(ind1-(w_ker_2D[ind1]).w))&&(csi<R))
out.at<uchar>(j-wm,i-wm)=(uchar) floor(IMG[(M+2*wm+1)*j+i]+0.5);
}
Mat out_cropped=out(Range(top,N-1-bottom),Range(left,M-1-right));
return out_cropped;
}
LogPolar_Overlapping::~LogPolar_Overlapping()
{
}
//----------------------------------------adjacent---------------------------------------
LogPolar_Adjacent::LogPolar_Adjacent(int w, int h, Point2i center, int _R, double _ro0, double smin, int full, int _s, int sp)
{
if ( (center.x!=w/2 || center.y!=h/2) && full==0) full=1;
if (center.x<0) center.x=0;
if (center.y<0) center.y=0;
if (center.x>=w) center.x=w-1;
if (center.y>=h) center.y=h-1;
if (full){
int rtmp;
if (center.x<=w/2 && center.y>=h/2)
rtmp=(int)std::sqrt((float)center.y*center.y + (float)(w-center.x)*(w-center.x));
else if (center.x>=w/2 && center.y>=h/2)
rtmp=(int)std::sqrt((float)center.y*center.y + (float)center.x*center.x);
else if (center.x>=w/2 && center.y<=h/2)
rtmp=(int)std::sqrt((float)(h-center.y)*(h-center.y) + (float)center.x*center.x);
else //if (center.x<=w/2 && center.y<=h/2)
rtmp=(int)std::sqrt((float)(h-center.y)*(h-center.y) + (float)(w-center.x)*(w-center.x));
M=2*rtmp; N=2*rtmp;
top = M/2 - center.y;
bottom = M/2 - (h-center.y);
left = M/2 - center.x;
right = M/2 - (w - center.x);
}else{
top=bottom=left=right=0;
M=w; N=h;
}
if (sp){
int jc=M/2-1, ic=N/2-1;
int _romax=std::min(ic, jc);
double _a=std::exp(std::log((double)(_romax/2-1)/(double)ro0)/(double)R);
S=(int) floor(2*CV_PI/(_a-1)+0.5);
}
create_map(M, N, _R, _s, _ro0, smin);
}
void LogPolar_Adjacent::create_map(int _M, int _n, int _R, int _s, double _ro0, double smin)
{
M=_M;
N=_n;
R=_R;
S=_s;
ro0=_ro0;
romax=std::min(M/2.0, N/2.0);
a=std::exp(std::log(romax/ro0)/(double)R);
q=S/(2*CV_PI);
A.resize(R*S);
L.resize(M*N);
for(int i=0; i<R*S; i++)
A[i]=0;
double xc=M/2.0, yc=N/2.0;
for(int j=0; j<N; j++)
for(int i=0; i<M; i++)
{
double x=i+0.5-xc, y=j+0.5-yc;
subdivide_recursively(x, y, i, j, 1, smin);
}
}
void LogPolar_Adjacent::subdivide_recursively(double x, double y, int i, int j, double length, double smin)
{
if(length<=smin)
{
int u, v;
if(get_uv(x, y, u, v))
{
pixel p;
p.u=u;
p.v=v;
p.a=length*length;
L[M*j+i].push_back(p);
A[v*R+u]+=length*length;
}
}
if(length>smin)
{
double xs[4], ys[4];
int us[4], vs[4];
xs[0]=xs[3]=x+length/4.0;
xs[1]=xs[2]=x-length/4.0;
ys[1]=ys[0]=y+length/4.0;
ys[2]=ys[3]=y-length/4.0;
for(int z=0; z<4; z++)
get_uv(xs[z], ys[z], us[z], vs[z]);
bool c=true;
for(int w=1; w<4; w++)
{
if(us[w]!=us[w-1])
c=false;
if(vs[w]!=vs[w-1])
c=false;
}
if(c)
{
if(us[0]!=-1)
{
pixel p;
p.u=us[0];
p.v=vs[0];
p.a=length*length;
L[M*j+i].push_back(p);
A[vs[0]*R+us[0]]+=length*length;
}
}
else
{
for(int z=0; z<4; z++)
if(us[z]!=-1)
subdivide_recursively(xs[z], ys[z], i, j, length/2.0, smin);
}
}
}
const Mat LogPolar_Adjacent::to_cortical(const Mat &source)
{
Mat source_border;
copyMakeBorder(source,source_border,top,bottom,left,right,BORDER_CONSTANT,Scalar(0));
std::vector<double> map(R*S, 0.);
for(int j=0; j<N; j++)
for(int i=0; i<M; i++)
{
for(size_t z=0; z<(L[M*j+i]).size(); z++)
{
map[R*((L[M*j+i])[z].v)+((L[M*j+i])[z].u)]+=((L[M*j+i])[z].a)*(source_border.at<uchar>(j,i));
}
}
for(int i=0; i<R*S; i++)
map[i]/=A[i];
Mat out(S,R,CV_8UC1,Scalar(0));
for(int i=0; i<S; i++)
for(int j=0;j<R;j++)
out.at<uchar>(i,j)=(uchar) floor(map[i*R+j]+0.5);
return out;
}
const Mat LogPolar_Adjacent::to_cartesian(const Mat &source)
{
std::vector<double> map(M*N, 0.);
for(int j=0; j<N; j++)
for(int i=0; i<M; i++)
{
for(size_t z=0; z<(L[M*j+i]).size(); z++)
{
map[M*j+i]+=(L[M*j+i])[z].a*source.at<uchar>((L[M*j+i])[z].v,(L[M*j+i])[z].u);
}
}
Mat out(N,M,CV_8UC1,Scalar(0));
for(int i=0; i<N; i++)
for(int j=0; j<M; j++)
out.at<uchar>(i,j)=(uchar) floor(map[i*M+j]+0.5);
Mat out_cropped=out(Range(top,N-1-bottom),Range(left,M-1-right));
return out_cropped;
}
bool LogPolar_Adjacent::get_uv(double x, double y, int&u, int&v)
{
double ro=std::sqrt(x*x+y*y), theta;
if(x>0)
theta=atan(y/x);
else
theta=atan(y/x)+CV_PI;
if(ro<ro0||ro>romax)
{
u=-1;
v=-1;
return false;
}
else
{
u= (int) floor(std::log(ro/ro0)/std::log(a));
if(theta>=0)
v= (int) floor(q*theta);
else
v= (int) floor(q*(theta+2*CV_PI));
return true;
}
}
LogPolar_Adjacent::~LogPolar_Adjacent()
{
}
}

344
modules/contrib/src/octree.cpp
View File

@@ -1,344 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "precomp.hpp"
#include <limits>
namespace
{
using namespace cv;
const size_t MAX_STACK_SIZE = 255;
const size_t MAX_LEAFS = 8;
bool checkIfNodeOutsideSphere(const Octree::Node& node, const Point3f& c, float r)
{
if (node.x_max < (c.x - r) || node.y_max < (c.y - r) || node.z_max < (c.z - r))
return true;
if ((c.x + r) < node.x_min || (c.y + r) < node.y_min || (c.z + r) < node.z_min)
return true;
return false;
}
bool checkIfNodeInsideSphere(const Octree::Node& node, const Point3f& c, float r)
{
r *= r;
float d2_xmin = (node.x_min - c.x) * (node.x_min - c.x);
float d2_ymin = (node.y_min - c.y) * (node.y_min - c.y);
float d2_zmin = (node.z_min - c.z) * (node.z_min - c.z);
if (d2_xmin + d2_ymin + d2_zmin > r)
return false;
float d2_zmax = (node.z_max - c.z) * (node.z_max - c.z);
if (d2_xmin + d2_ymin + d2_zmax > r)
return false;
float d2_ymax = (node.y_max - c.y) * (node.y_max - c.y);
if (d2_xmin + d2_ymax + d2_zmin > r)
return false;
if (d2_xmin + d2_ymax + d2_zmax > r)
return false;
float d2_xmax = (node.x_max - c.x) * (node.x_max - c.x);
if (d2_xmax + d2_ymin + d2_zmin > r)
return false;
if (d2_xmax + d2_ymin + d2_zmax > r)
return false;
if (d2_xmax + d2_ymax + d2_zmin > r)
return false;
if (d2_xmax + d2_ymax + d2_zmax > r)
return false;
return true;
}
void fillMinMax(const std::vector<Point3f>& points, Octree::Node& node)
{
node.x_max = node.y_max = node.z_max = std::numeric_limits<float>::min();
node.x_min = node.y_min = node.z_min = std::numeric_limits<float>::max();
for (size_t i = 0; i < points.size(); ++i)
{
const Point3f& point = points[i];
if (node.x_max < point.x)
node.x_max = point.x;
if (node.y_max < point.y)
node.y_max = point.y;
if (node.z_max < point.z)
node.z_max = point.z;
if (node.x_min > point.x)
node.x_min = point.x;
if (node.y_min > point.y)
node.y_min = point.y;
if (node.z_min > point.z)
node.z_min = point.z;
}
}
size_t findSubboxForPoint(const Point3f& point, const Octree::Node& node)
{
size_t ind_x = point.x < (node.x_max + node.x_min) / 2 ? 0 : 1;
size_t ind_y = point.y < (node.y_max + node.y_min) / 2 ? 0 : 1;
size_t ind_z = point.z < (node.z_max + node.z_min) / 2 ? 0 : 1;
return (ind_x << 2) + (ind_y << 1) + (ind_z << 0);
}
void initChildBox(const Octree::Node& parent, size_t boxIndex, Octree::Node& child)
{
child.x_min = child.x_max = (parent.x_max + parent.x_min) / 2;
child.y_min = child.y_max = (parent.y_max + parent.y_min) / 2;
child.z_min = child.z_max = (parent.z_max + parent.z_min) / 2;
if ((boxIndex >> 0) & 1)
child.z_max = parent.z_max;
else
child.z_min = parent.z_min;
if ((boxIndex >> 1) & 1)
child.y_max = parent.y_max;
else
child.y_min = parent.y_min;
if ((boxIndex >> 2) & 1)
child.x_max = parent.x_max;
else
child.x_min = parent.x_min;
}
}//namespace
////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////// Octree //////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////////////
namespace cv
{
Octree::Octree()
{
}
Octree::Octree(const std::vector<Point3f>& points3d, int maxLevels, int _minPoints)
{
buildTree(points3d, maxLevels, _minPoints);
}
Octree::~Octree()
{
}
void Octree::getPointsWithinSphere(const Point3f& center, float radius, std::vector<Point3f>& out) const
{
out.clear();
if (nodes.empty())
return;
int stack[MAX_STACK_SIZE];
int pos = 0;
stack[pos] = 0;
while (pos >= 0)
{
const Node& cur = nodes[stack[pos--]];
if (checkIfNodeOutsideSphere(cur, center, radius))
continue;
if (checkIfNodeInsideSphere(cur, center, radius))
{
size_t sz = out.size();
out.resize(sz + cur.end - cur.begin);
for (int i = cur.begin; i < cur.end; ++i)
out[sz++] = points[i];
continue;
}
if (cur.isLeaf)
{
double r2 = radius * radius;
size_t sz = out.size();
out.resize(sz + (cur.end - cur.begin));
for (int i = cur.begin; i < cur.end; ++i)
{
const Point3f& point = points[i];
double dx = (point.x - center.x);
double dy = (point.y - center.y);
double dz = (point.z - center.z);
double dist2 = dx * dx + dy * dy + dz * dz;
if (dist2 < r2)
out[sz++] = point;
};
out.resize(sz);
continue;
}
if (cur.children[0])
stack[++pos] = cur.children[0];
if (cur.children[1])
stack[++pos] = cur.children[1];
if (cur.children[2])
stack[++pos] = cur.children[2];
if (cur.children[3])
stack[++pos] = cur.children[3];
if (cur.children[4])
stack[++pos] = cur.children[4];
if (cur.children[5])
stack[++pos] = cur.children[5];
if (cur.children[6])
stack[++pos] = cur.children[6];
if (cur.children[7])
stack[++pos] = cur.children[7];
}
}
void Octree::buildTree(const std::vector<Point3f>& points3d, int maxLevels, int _minPoints)
{
CV_Assert((size_t)maxLevels * 8 < MAX_STACK_SIZE);
points.resize(points3d.size());
std::copy(points3d.begin(), points3d.end(), points.begin());
minPoints = _minPoints;
nodes.clear();
nodes.push_back(Node());
Node& root = nodes[0];
fillMinMax(points, root);
root.isLeaf = true;
root.maxLevels = maxLevels;
root.begin = 0;
root.end = (int)points.size();
for (size_t i = 0; i < MAX_LEAFS; i++)
root.children[i] = 0;
if (maxLevels != 1 && (root.end - root.begin) > _minPoints)
{
root.isLeaf = false;
buildNext(0);
}
}
void Octree::buildNext(size_t nodeInd)
{
size_t size = nodes[nodeInd].end - nodes[nodeInd].begin;
std::vector<size_t> boxBorders(MAX_LEAFS+1, 0);
std::vector<size_t> boxIndices(size);
std::vector<Point3f> tempPoints(size);
for (int i = nodes[nodeInd].begin, j = 0; i < nodes[nodeInd].end; ++i, ++j)
{
const Point3f& p = points[i];
size_t subboxInd = findSubboxForPoint(p, nodes[nodeInd]);
boxBorders[subboxInd+1]++;
boxIndices[j] = subboxInd;
tempPoints[j] = p;
}
for (size_t i = 1; i < boxBorders.size(); ++i)
boxBorders[i] += boxBorders[i-1];
std::vector<size_t> writeInds(boxBorders.begin(), boxBorders.end());
for (size_t i = 0; i < size; ++i)
{
size_t boxIndex = boxIndices[i];
Point3f& curPoint = tempPoints[i];
size_t copyTo = nodes[nodeInd].begin + writeInds[boxIndex]++;
points[copyTo] = curPoint;
}
for (size_t i = 0; i < MAX_LEAFS; ++i)
{
if (boxBorders[i] == boxBorders[i+1])
continue;
nodes.push_back(Node());
Node& child = nodes.back();
initChildBox(nodes[nodeInd], i, child);
child.isLeaf = true;
child.maxLevels = nodes[nodeInd].maxLevels - 1;
child.begin = nodes[nodeInd].begin + (int)boxBorders[i+0];
child.end = nodes[nodeInd].begin + (int)boxBorders[i+1];
for (size_t k = 0; k < MAX_LEAFS; k++)
child.children[k] = 0;
nodes[nodeInd].children[i] = (int)(nodes.size() - 1);
if (child.maxLevels != 1 && (child.end - child.begin) > minPoints)
{
child.isLeaf = false;
buildNext(nodes.size() - 1);
}
}
}
}

778
modules/contrib/src/openfabmap.cpp
View File

@@ -1,778 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
// This file originates from the openFABMAP project:
// [http://code.google.com/p/openfabmap/]
//
// For published work which uses all or part of OpenFABMAP, please cite:
// [http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6224843]
//
// Original Algorithm by Mark Cummins and Paul Newman:
// [http://ijr.sagepub.com/content/27/6/647.short]
// [http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5613942]
// [http://ijr.sagepub.com/content/30/9/1100.abstract]
//
// License Agreement
//
// Copyright (C) 2012 Arren Glover [aj.glover@qut.edu.au] and
// Will Maddern [w.maddern@qut.edu.au], all rights reserved.
//
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "precomp.hpp"
#include "opencv2/contrib/openfabmap.hpp"
/*
Calculate the sum of two log likelihoods
*/
namespace cv {
namespace of2 {
static double logsumexp(double a, double b) {
return a > b ? std::log(1 + std::exp(b - a)) + a : std::log(1 + std::exp(a - b)) + b;
}
FabMap::FabMap(const Mat& _clTree, double _PzGe,
double _PzGNe, int _flags, int _numSamples) :
clTree(_clTree), PzGe(_PzGe), PzGNe(_PzGNe), flags(
_flags), numSamples(_numSamples) {
CV_Assert(flags & MEAN_FIELD || flags & SAMPLED);
CV_Assert(flags & NAIVE_BAYES || flags & CHOW_LIU);
if (flags & NAIVE_BAYES) {
PzGL = &FabMap::PzqGL;
} else {
PzGL = &FabMap::PzqGzpqL;
}
//check for a valid Chow-Liu tree
CV_Assert(clTree.type() == CV_64FC1);
cv::checkRange(clTree.row(0), false, NULL, 0, clTree.cols);
cv::checkRange(clTree.row(1), false, NULL, DBL_MIN, 1);
cv::checkRange(clTree.row(2), false, NULL, DBL_MIN, 1);
cv::checkRange(clTree.row(3), false, NULL, DBL_MIN, 1);
// TODO: Add default values for member variables
Pnew = 0.9;
sFactor = 0.99;
mBias = 0.5;
}
FabMap::~FabMap() {
}
const std::vector<cv::Mat>& FabMap::getTrainingImgDescriptors() const {
return trainingImgDescriptors;
}
const std::vector<cv::Mat>& FabMap::getTestImgDescriptors() const {
return testImgDescriptors;
}
void FabMap::addTraining(const Mat& queryImgDescriptor) {
CV_Assert(!queryImgDescriptor.empty());
std::vector<Mat> queryImgDescriptors;
for (int i = 0; i < queryImgDescriptor.rows; i++) {
queryImgDescriptors.push_back(queryImgDescriptor.row(i));
}
addTraining(queryImgDescriptors);
}
void FabMap::addTraining(const std::vector<Mat>& queryImgDescriptors) {
for (size_t i = 0; i < queryImgDescriptors.size(); i++) {
CV_Assert(!queryImgDescriptors[i].empty());
CV_Assert(queryImgDescriptors[i].rows == 1);
CV_Assert(queryImgDescriptors[i].cols == clTree.cols);
CV_Assert(queryImgDescriptors[i].type() == CV_32F);
trainingImgDescriptors.push_back(queryImgDescriptors[i]);
}
}
void FabMap::add(const cv::Mat& queryImgDescriptor) {
CV_Assert(!queryImgDescriptor.empty());
std::vector<Mat> queryImgDescriptors;
for (int i = 0; i < queryImgDescriptor.rows; i++) {
queryImgDescriptors.push_back(queryImgDescriptor.row(i));
}
add(queryImgDescriptors);
}
void FabMap::add(const std::vector<cv::Mat>& queryImgDescriptors) {
for (size_t i = 0; i < queryImgDescriptors.size(); i++) {
CV_Assert(!queryImgDescriptors[i].empty());
CV_Assert(queryImgDescriptors[i].rows == 1);
CV_Assert(queryImgDescriptors[i].cols == clTree.cols);
CV_Assert(queryImgDescriptors[i].type() == CV_32F);
testImgDescriptors.push_back(queryImgDescriptors[i]);
}
}
void FabMap::compare(const Mat& queryImgDescriptor,
std::vector<IMatch>& matches, bool addQuery,
const Mat& mask) {
CV_Assert(!queryImgDescriptor.empty());
std::vector<Mat> queryImgDescriptors;
for (int i = 0; i < queryImgDescriptor.rows; i++) {
queryImgDescriptors.push_back(queryImgDescriptor.row(i));
}
compare(queryImgDescriptors,matches,addQuery,mask);
}
void FabMap::compare(const Mat& queryImgDescriptor,
const Mat& testImgDescriptor, std::vector<IMatch>& matches,
const Mat& mask) {
CV_Assert(!queryImgDescriptor.empty());
std::vector<Mat> queryImgDescriptors;
for (int i = 0; i < queryImgDescriptor.rows; i++) {
queryImgDescriptors.push_back(queryImgDescriptor.row(i));
}
CV_Assert(!testImgDescriptor.empty());
std::vector<Mat> _testImgDescriptors;
for (int i = 0; i < testImgDescriptor.rows; i++) {
_testImgDescriptors.push_back(testImgDescriptor.row(i));
}
compare(queryImgDescriptors,_testImgDescriptors,matches,mask);
}
void FabMap::compare(const Mat& queryImgDescriptor,
const std::vector<Mat>& _testImgDescriptors,
std::vector<IMatch>& matches, const Mat& mask) {
CV_Assert(!queryImgDescriptor.empty());
std::vector<Mat> queryImgDescriptors;
for (int i = 0; i < queryImgDescriptor.rows; i++) {
queryImgDescriptors.push_back(queryImgDescriptor.row(i));
}
compare(queryImgDescriptors,_testImgDescriptors,matches,mask);
}
void FabMap::compare(const std::vector<Mat>& queryImgDescriptors,
std::vector<IMatch>& matches, bool addQuery, const Mat& /*mask*/) {
// TODO: add first query if empty (is this necessary)
for (size_t i = 0; i < queryImgDescriptors.size(); i++) {
CV_Assert(!queryImgDescriptors[i].empty());
CV_Assert(queryImgDescriptors[i].rows == 1);
CV_Assert(queryImgDescriptors[i].cols == clTree.cols);
CV_Assert(queryImgDescriptors[i].type() == CV_32F);
// TODO: add mask
compareImgDescriptor(queryImgDescriptors[i],
(int)i, testImgDescriptors, matches);
if (addQuery)
add(queryImgDescriptors[i]);
}
}
void FabMap::compare(const std::vector<Mat>& queryImgDescriptors,
const std::vector<Mat>& _testImgDescriptors,
std::vector<IMatch>& matches, const Mat& /*mask*/) {
CV_Assert(!(flags & MOTION_MODEL));
for (size_t i = 0; i < _testImgDescriptors.size(); i++) {
CV_Assert(!_testImgDescriptors[i].empty());
CV_Assert(_testImgDescriptors[i].rows == 1);
CV_Assert(_testImgDescriptors[i].cols == clTree.cols);
CV_Assert(_testImgDescriptors[i].type() == CV_32F);
}
for (size_t i = 0; i < queryImgDescriptors.size(); i++) {
CV_Assert(!queryImgDescriptors[i].empty());
CV_Assert(queryImgDescriptors[i].rows == 1);
CV_Assert(queryImgDescriptors[i].cols == clTree.cols);
CV_Assert(queryImgDescriptors[i].type() == CV_32F);
// TODO: add mask
compareImgDescriptor(queryImgDescriptors[i],
(int)i, _testImgDescriptors, matches);
}
}
void FabMap::compareImgDescriptor(const Mat& queryImgDescriptor,
int queryIndex, const std::vector<Mat>& _testImgDescriptors,
std::vector<IMatch>& matches) {
std::vector<IMatch> queryMatches;
queryMatches.push_back(IMatch(queryIndex,-1,
getNewPlaceLikelihood(queryImgDescriptor),0));
getLikelihoods(queryImgDescriptor,_testImgDescriptors,queryMatches);
normaliseDistribution(queryMatches);
for (size_t j = 1; j < queryMatches.size(); j++) {
queryMatches[j].queryIdx = queryIndex;
}
matches.insert(matches.end(), queryMatches.begin(), queryMatches.end());
}
void FabMap::getLikelihoods(const Mat& /*queryImgDescriptor*/,
const std::vector<Mat>& /*testImgDescriptors*/, std::vector<IMatch>& /*matches*/) {
}
double FabMap::getNewPlaceLikelihood(const Mat& queryImgDescriptor) {
if (flags & MEAN_FIELD) {
double logP = 0;
bool zq, zpq;
if(flags & NAIVE_BAYES) {
for (int q = 0; q < clTree.cols; q++) {
zq = queryImgDescriptor.at<float>(0,q) > 0;
logP += std::log(Pzq(q, false) * PzqGeq(zq, false) +
Pzq(q, true) * PzqGeq(zq, true));
}
} else {
for (int q = 0; q < clTree.cols; q++) {
zq = queryImgDescriptor.at<float>(0,q) > 0;
zpq = queryImgDescriptor.at<float>(0,pq(q)) > 0;
double alpha, beta, p;
alpha = Pzq(q, zq) * PzqGeq(!zq, false) * PzqGzpq(q, !zq, zpq);
beta = Pzq(q, !zq) * PzqGeq(zq, false) * PzqGzpq(q, zq, zpq);
p = Pzq(q, false) * beta / (alpha + beta);
alpha = Pzq(q, zq) * PzqGeq(!zq, true) * PzqGzpq(q, !zq, zpq);
beta = Pzq(q, !zq) * PzqGeq(zq, true) * PzqGzpq(q, zq, zpq);
p += Pzq(q, true) * beta / (alpha + beta);
logP += std::log(p);
}
}
return logP;
}
if (flags & SAMPLED) {
CV_Assert(!trainingImgDescriptors.empty());
CV_Assert(numSamples > 0);
std::vector<Mat> sampledImgDescriptors;
// TODO: this method can result in the same sample being added
// multiple times. Is this desired?
for (int i = 0; i < numSamples; i++) {
int index = rand() % trainingImgDescriptors.size();
sampledImgDescriptors.push_back(trainingImgDescriptors[index]);
}
std::vector<IMatch> matches;
getLikelihoods(queryImgDescriptor,sampledImgDescriptors,matches);
double averageLogLikelihood = -DBL_MAX + matches.front().likelihood + 1;
for (int i = 0; i < numSamples; i++) {
averageLogLikelihood =
logsumexp(matches[i].likelihood, averageLogLikelihood);
}
return averageLogLikelihood - std::log((double)numSamples);
}
return 0;
}
void FabMap::normaliseDistribution(std::vector<IMatch>& matches) {
CV_Assert(!matches.empty());
if (flags & MOTION_MODEL) {
matches[0].match = matches[0].likelihood + std::log(Pnew);
if (priorMatches.size() > 2) {
matches[1].match = matches[1].likelihood;
matches[1].match += std::log(
(2 * (1-mBias) * priorMatches[1].match +
priorMatches[1].match +
2 * mBias * priorMatches[2].match) / 3);
for (size_t i = 2; i < priorMatches.size()-1; i++) {
matches[i].match = matches[i].likelihood;
matches[i].match += std::log(
(2 * (1-mBias) * priorMatches[i-1].match +
priorMatches[i].match +
2 * mBias * priorMatches[i+1].match)/3);
}
matches[priorMatches.size()-1].match =
matches[priorMatches.size()-1].likelihood;
matches[priorMatches.size()-1].match += std::log(
(2 * (1-mBias) * priorMatches[priorMatches.size()-2].match +
priorMatches[priorMatches.size()-1].match +
2 * mBias * priorMatches[priorMatches.size()-1].match)/3);
for(size_t i = priorMatches.size(); i < matches.size(); i++) {
matches[i].match = matches[i].likelihood;
}
} else {
for(size_t i = 1; i < matches.size(); i++) {
matches[i].match = matches[i].likelihood;
}
}
double logsum = -DBL_MAX + matches.front().match + 1;
//calculate the normalising constant
for (size_t i = 0; i < matches.size(); i++) {
logsum = logsumexp(logsum, matches[i].match);
}
//normalise
for (size_t i = 0; i < matches.size(); i++) {
matches[i].match = std::exp(matches[i].match - logsum);
}
//smooth final probabilities
for (size_t i = 0; i < matches.size(); i++) {
matches[i].match = sFactor*matches[i].match +
(1 - sFactor)/matches.size();
}
//update our location priors
priorMatches = matches;
} else {
double logsum = -DBL_MAX + matches.front().likelihood + 1;
for (size_t i = 0; i < matches.size(); i++) {
logsum = logsumexp(logsum, matches[i].likelihood);
}
for (size_t i = 0; i < matches.size(); i++) {
matches[i].match = std::exp(matches[i].likelihood - logsum);
}
for (size_t i = 0; i < matches.size(); i++) {
matches[i].match = sFactor*matches[i].match +
(1 - sFactor)/matches.size();
}
}
}
int FabMap::pq(int q) {
return (int)clTree.at<double>(0,q);
}
double FabMap::Pzq(int q, bool zq) {
return (zq) ? clTree.at<double>(1,q) : 1 - clTree.at<double>(1,q);
}
double FabMap::PzqGzpq(int q, bool zq, bool zpq) {
if (zpq) {
return (zq) ? clTree.at<double>(2,q) : 1 - clTree.at<double>(2,q);
} else {
return (zq) ? clTree.at<double>(3,q) : 1 - clTree.at<double>(3,q);
}
}
double FabMap::PzqGeq(bool zq, bool eq) {
if (eq) {
return (zq) ? PzGe : 1 - PzGe;
} else {
return (zq) ? PzGNe : 1 - PzGNe;
}
}
double FabMap::PeqGL(int q, bool Lzq, bool eq) {
double alpha, beta;
alpha = PzqGeq(Lzq, true) * Pzq(q, true);
beta = PzqGeq(Lzq, false) * Pzq(q, false);
if (eq) {
return alpha / (alpha + beta);
} else {
return 1 - (alpha / (alpha + beta));
}
}
double FabMap::PzqGL(int q, bool zq, bool /*zpq*/, bool Lzq) {
return PeqGL(q, Lzq, false) * PzqGeq(zq, false) +
PeqGL(q, Lzq, true) * PzqGeq(zq, true);
}
double FabMap::PzqGzpqL(int q, bool zq, bool zpq, bool Lzq) {
double p;
double alpha, beta;
alpha = Pzq(q, zq) * PzqGeq(!zq, false) * PzqGzpq(q, !zq, zpq);
beta = Pzq(q, !zq) * PzqGeq( zq, false) * PzqGzpq(q, zq, zpq);
p = PeqGL(q, Lzq, false) * beta / (alpha + beta);
alpha = Pzq(q, zq) * PzqGeq(!zq, true) * PzqGzpq(q, !zq, zpq);
beta = Pzq(q, !zq) * PzqGeq( zq, true) * PzqGzpq(q, zq, zpq);
p += PeqGL(q, Lzq, true) * beta / (alpha + beta);
return p;
}
FabMap1::FabMap1(const Mat& _clTree, double _PzGe, double _PzGNe, int _flags,
int _numSamples) : FabMap(_clTree, _PzGe, _PzGNe, _flags,
_numSamples) {
}
FabMap1::~FabMap1() {
}
void FabMap1::getLikelihoods(const Mat& queryImgDescriptor,
const std::vector<Mat>& testImageDescriptors, std::vector<IMatch>& matches) {
for (size_t i = 0; i < testImageDescriptors.size(); i++) {
bool zq, zpq, Lzq;
double logP = 0;
for (int q = 0; q < clTree.cols; q++) {
zq = queryImgDescriptor.at<float>(0,q) > 0;
zpq = queryImgDescriptor.at<float>(0,pq(q)) > 0;
Lzq = testImageDescriptors[i].at<float>(0,q) > 0;
logP += std::log((this->*PzGL)(q, zq, zpq, Lzq));
}
matches.push_back(IMatch(0,(int)i,logP,0));
}
}
FabMapLUT::FabMapLUT(const Mat& _clTree, double _PzGe, double _PzGNe,
int _flags, int _numSamples, int _precision) :
FabMap(_clTree, _PzGe, _PzGNe, _flags, _numSamples), precision(_precision) {
int nWords = clTree.cols;
double precFactor = (double)std::pow(10.0, precision);
table = new int[nWords][8];
for (int q = 0; q < nWords; q++) {
for (unsigned char i = 0; i < 8; i++) {
bool Lzq = (bool) ((i >> 2) & 0x01);
bool zq = (bool) ((i >> 1) & 0x01);
bool zpq = (bool) (i & 1);
table[q][i] = -(int)(std::log((this->*PzGL)(q, zq, zpq, Lzq))
* precFactor);
}
}
}
FabMapLUT::~FabMapLUT() {
delete[] table;
}
void FabMapLUT::getLikelihoods(const Mat& queryImgDescriptor,
const std::vector<Mat>& testImageDescriptors, std::vector<IMatch>& matches) {
double precFactor = (double)std::pow(10.0, -precision);
for (size_t i = 0; i < testImageDescriptors.size(); i++) {
unsigned long long int logP = 0;
for (int q = 0; q < clTree.cols; q++) {
logP += table[q][(queryImgDescriptor.at<float>(0,pq(q)) > 0) +
((queryImgDescriptor.at<float>(0, q) > 0) << 1) +
((testImageDescriptors[i].at<float>(0,q) > 0) << 2)];
}
matches.push_back(IMatch(0,(int)i,-precFactor*(double)logP,0));
}
}
FabMapFBO::FabMapFBO(const Mat& _clTree, double _PzGe, double _PzGNe,
int _flags, int _numSamples, double _rejectionThreshold,
double _PsGd, int _bisectionStart, int _bisectionIts) :
FabMap(_clTree, _PzGe, _PzGNe, _flags, _numSamples), PsGd(_PsGd),
rejectionThreshold(_rejectionThreshold), bisectionStart(_bisectionStart),
bisectionIts(_bisectionIts) {
}
FabMapFBO::~FabMapFBO() {
}
void FabMapFBO::getLikelihoods(const Mat& queryImgDescriptor,
const std::vector<Mat>& testImageDescriptors, std::vector<IMatch>& matches) {
std::multiset<WordStats> wordData;
setWordStatistics(queryImgDescriptor, wordData);
std::vector<int> matchIndices;
std::vector<IMatch> queryMatches;
for (size_t i = 0; i < testImageDescriptors.size(); i++) {
queryMatches.push_back(IMatch(0,(int)i,0,0));
matchIndices.push_back((int)i);
}
double currBest = -DBL_MAX;
double bailedOut = DBL_MAX;
for (std::multiset<WordStats>::iterator wordIter = wordData.begin();
wordIter != wordData.end(); wordIter++) {
bool zq = queryImgDescriptor.at<float>(0,wordIter->q) > 0;
bool zpq = queryImgDescriptor.at<float>(0,pq(wordIter->q)) > 0;
currBest = -DBL_MAX;
for (size_t i = 0; i < matchIndices.size(); i++) {
bool Lzq =
testImageDescriptors[matchIndices[i]].at<float>(0,wordIter->q) > 0;
queryMatches[matchIndices[i]].likelihood +=
std::log((this->*PzGL)(wordIter->q,zq,zpq,Lzq));
currBest =
std::max(queryMatches[matchIndices[i]].likelihood, currBest);
}
if (matchIndices.size() == 1)
continue;
double delta = std::max(limitbisection(wordIter->V, wordIter->M),
-std::log(rejectionThreshold));
std::vector<int>::iterator matchIter = matchIndices.begin();
while (matchIter != matchIndices.end()) {
if (currBest - queryMatches[*matchIter].likelihood > delta) {
queryMatches[*matchIter].likelihood = bailedOut;
matchIter = matchIndices.erase(matchIter);
} else {
matchIter++;
}
}
}
for (size_t i = 0; i < queryMatches.size(); i++) {
if (queryMatches[i].likelihood == bailedOut) {
queryMatches[i].likelihood = currBest + std::log(rejectionThreshold);
}
}
matches.insert(matches.end(), queryMatches.begin(), queryMatches.end());
}
void FabMapFBO::setWordStatistics(const Mat& queryImgDescriptor,
std::multiset<WordStats>& wordData) {
//words are sorted according to information = -ln(P(zq|zpq))
//in non-log format this is lowest probability first
for (int q = 0; q < clTree.cols; q++) {
wordData.insert(WordStats(q,PzqGzpq(q,
queryImgDescriptor.at<float>(0,q) > 0,
queryImgDescriptor.at<float>(0,pq(q)) > 0)));
}
double d = 0, V = 0, M = 0;
bool zq, zpq;
for (std::multiset<WordStats>::reverse_iterator wordIter =
wordData.rbegin();
wordIter != wordData.rend(); wordIter++) {
zq = queryImgDescriptor.at<float>(0,wordIter->q) > 0;
zpq = queryImgDescriptor.at<float>(0,pq(wordIter->q)) > 0;
d = std::log((this->*PzGL)(wordIter->q, zq, zpq, true)) -
std::log((this->*PzGL)(wordIter->q, zq, zpq, false));
V += std::pow(d, 2.0) * 2 *
(Pzq(wordIter->q, true) - std::pow(Pzq(wordIter->q, true), 2.0));
M = std::max(M, fabs(d));
wordIter->V = V;
wordIter->M = M;
}
}
double FabMapFBO::limitbisection(double v, double m) {
double midpoint, left_val, mid_val;
double left = 0, right = bisectionStart;
left_val = bennettInequality(v, m, left) - PsGd;
for(int i = 0; i < bisectionIts; i++) {
midpoint = (left + right)*0.5;
mid_val = bennettInequality(v, m, midpoint)- PsGd;
if(left_val * mid_val > 0) {
left = midpoint;
left_val = mid_val;
} else {
right = midpoint;
}
}
return (right + left) * 0.5;
}
double FabMapFBO::bennettInequality(double v, double m, double delta) {
double DMonV = delta * m / v;
double f_delta = std::log(DMonV + std::sqrt(std::pow(DMonV, 2.0) + 1));
return std::exp((v / std::pow(m, 2.0))*(cosh(f_delta) - 1 - DMonV * f_delta));
}
bool FabMapFBO::compInfo(const WordStats& first, const WordStats& second) {
return first.info < second.info;
}
FabMap2::FabMap2(const Mat& _clTree, double _PzGe, double _PzGNe,
int _flags) :
FabMap(_clTree, _PzGe, _PzGNe, _flags) {
CV_Assert(flags & SAMPLED);
children.resize(clTree.cols);
for (int q = 0; q < clTree.cols; q++) {
d1.push_back(std::log((this->*PzGL)(q, false, false, true) /
(this->*PzGL)(q, false, false, false)));
d2.push_back(std::log((this->*PzGL)(q, false, true, true) /
(this->*PzGL)(q, false, true, false)) - d1[q]);
d3.push_back(std::log((this->*PzGL)(q, true, false, true) /
(this->*PzGL)(q, true, false, false))- d1[q]);
d4.push_back(std::log((this->*PzGL)(q, true, true, true) /
(this->*PzGL)(q, true, true, false))- d1[q]);
children[pq(q)].push_back(q);
}
}
FabMap2::~FabMap2() {
}
void FabMap2::addTraining(const std::vector<Mat>& queryImgDescriptors) {
for (size_t i = 0; i < queryImgDescriptors.size(); i++) {
CV_Assert(!queryImgDescriptors[i].empty());
CV_Assert(queryImgDescriptors[i].rows == 1);
CV_Assert(queryImgDescriptors[i].cols == clTree.cols);
CV_Assert(queryImgDescriptors[i].type() == CV_32F);
trainingImgDescriptors.push_back(queryImgDescriptors[i]);
addToIndex(queryImgDescriptors[i], trainingDefaults, trainingInvertedMap);
}
}
void FabMap2::add(const std::vector<Mat>& queryImgDescriptors) {
for (size_t i = 0; i < queryImgDescriptors.size(); i++) {
CV_Assert(!queryImgDescriptors[i].empty());
CV_Assert(queryImgDescriptors[i].rows == 1);
CV_Assert(queryImgDescriptors[i].cols == clTree.cols);
CV_Assert(queryImgDescriptors[i].type() == CV_32F);
testImgDescriptors.push_back(queryImgDescriptors[i]);
addToIndex(queryImgDescriptors[i], testDefaults, testInvertedMap);
}
}
void FabMap2::getLikelihoods(const Mat& queryImgDescriptor,
const std::vector<Mat>& testImageDescriptors, std::vector<IMatch>& matches) {
if (&testImageDescriptors == &testImgDescriptors) {
getIndexLikelihoods(queryImgDescriptor, testDefaults, testInvertedMap,
matches);
} else {
CV_Assert(!(flags & MOTION_MODEL));
std::vector<double> defaults;
std::map<int, std::vector<int> > invertedMap;
for (size_t i = 0; i < testImageDescriptors.size(); i++) {
addToIndex(testImageDescriptors[i],defaults,invertedMap);
}
getIndexLikelihoods(queryImgDescriptor, defaults, invertedMap, matches);
}
}
double FabMap2::getNewPlaceLikelihood(const Mat& queryImgDescriptor) {
CV_Assert(!trainingImgDescriptors.empty());
std::vector<IMatch> matches;
getIndexLikelihoods(queryImgDescriptor, trainingDefaults,
trainingInvertedMap, matches);
double averageLogLikelihood = -DBL_MAX + matches.front().likelihood + 1;
for (size_t i = 0; i < matches.size(); i++) {
averageLogLikelihood =
logsumexp(matches[i].likelihood, averageLogLikelihood);
}
return averageLogLikelihood - std::log((double)trainingDefaults.size());
}
void FabMap2::addToIndex(const Mat& queryImgDescriptor,
std::vector<double>& defaults,
std::map<int, std::vector<int> >& invertedMap) {
defaults.push_back(0);
for (int q = 0; q < clTree.cols; q++) {
if (queryImgDescriptor.at<float>(0,q) > 0) {
defaults.back() += d1[q];
invertedMap[q].push_back((int)defaults.size()-1);
}
}
}
void FabMap2::getIndexLikelihoods(const Mat& queryImgDescriptor,
std::vector<double>& defaults,
std::map<int, std::vector<int> >& invertedMap,
std::vector<IMatch>& matches) {
std::vector<int>::iterator LwithI, child;
std::vector<double> likelihoods = defaults;
for (int q = 0; q < clTree.cols; q++) {
if (queryImgDescriptor.at<float>(0,q) > 0) {
for (LwithI = invertedMap[q].begin();
LwithI != invertedMap[q].end(); LwithI++) {
if (queryImgDescriptor.at<float>(0,pq(q)) > 0) {
likelihoods[*LwithI] += d4[q];
} else {
likelihoods[*LwithI] += d3[q];
}
}
for (child = children[q].begin(); child != children[q].end();
child++) {
if (queryImgDescriptor.at<float>(0,*child) == 0) {
for (LwithI = invertedMap[*child].begin();
LwithI != invertedMap[*child].end(); LwithI++) {
likelihoods[*LwithI] += d2[*child];
}
}
}
}
}
for (size_t i = 0; i < likelihoods.size(); i++) {
matches.push_back(IMatch(0,(int)i,likelihoods[i],0));
}
}
}
}

83
modules/contrib/src/polyfit.cpp
View File

@@ -1,83 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
// This original code was written by
// Onkar Raut
// Graduate Student,
// University of North Carolina at Charlotte
#include "precomp.hpp"
typedef double polyfit_type;
void cv::polyfit(const Mat& src_x, const Mat& src_y, Mat& dst, int order)
{
const int wdepth = DataType<polyfit_type>::depth;
int npoints = src_x.checkVector(1);
int nypoints = src_y.checkVector(1);
CV_Assert(npoints == nypoints && npoints >= order+1);
Mat srcX = Mat_<polyfit_type>(src_x), srcY = Mat_<polyfit_type>(src_y);
Mat X = Mat::zeros(order + 1, npoints, wdepth);
polyfit_type* pSrcX = (polyfit_type*)srcX.data;
polyfit_type* pXData = (polyfit_type*)X.data;
int stepX = (int)(X.step/X.elemSize1());
for (int y = 0; y < order + 1; ++y)
{
for (int x = 0; x < npoints; ++x)
{
if (y == 0)
pXData[x] = 1;
else if (y == 1)
pXData[x + stepX] = pSrcX[x];
else pXData[x + y*stepX] = pSrcX[x]* pXData[x + (y-1)*stepX];
}
}
Mat A, b, w;
mulTransposed(X, A, false);
b = X*srcY;
solve(A, b, w, DECOMP_SVD);
w.convertTo(dst, std::max(std::max(src_x.depth(), src_y.depth()), CV_32F));
}

635
modules/contrib/src/rgbdodometry.cpp
View File

@@ -1,635 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "precomp.hpp"
#define SHOW_DEBUG_IMAGES 0
#include "opencv2/core.hpp"
#include "opencv2/calib3d.hpp"
#if SHOW_DEBUG_IMAGES
# include "opencv2/highgui.hpp"
#endif
#include <iostream>
#include <limits>
#if defined(HAVE_EIGEN) && EIGEN_WORLD_VERSION == 3
# ifdef ANDROID
template <typename Scalar> Scalar log2(Scalar v) { return std::log(v)/std::log(Scalar(2)); }
# endif
# if defined __GNUC__ && defined __APPLE__
# pragma GCC diagnostic ignored "-Wshadow"
# endif
# include <unsupported/Eigen/MatrixFunctions>
# include <Eigen/Dense>
#endif
using namespace cv;
inline static
void computeC_RigidBodyMotion( double* C, double dIdx, double dIdy, const Point3f& p3d, double fx, double fy )
{
double invz = 1. / p3d.z,
v0 = dIdx * fx * invz,
v1 = dIdy * fy * invz,
v2 = -(v0 * p3d.x + v1 * p3d.y) * invz;
C[0] = -p3d.z * v1 + p3d.y * v2;
C[1] = p3d.z * v0 - p3d.x * v2;
C[2] = -p3d.y * v0 + p3d.x * v1;
C[3] = v0;
C[4] = v1;
C[5] = v2;
}
inline static
void computeC_Rotation( double* C, double dIdx, double dIdy, const Point3f& p3d, double fx, double fy )
{
double invz = 1. / p3d.z,
v0 = dIdx * fx * invz,
v1 = dIdy * fy * invz,
v2 = -(v0 * p3d.x + v1 * p3d.y) * invz;
C[0] = -p3d.z * v1 + p3d.y * v2;
C[1] = p3d.z * v0 - p3d.x * v2;
C[2] = -p3d.y * v0 + p3d.x * v1;
}
inline static
void computeC_Translation( double* C, double dIdx, double dIdy, const Point3f& p3d, double fx, double fy )
{
double invz = 1. / p3d.z,
v0 = dIdx * fx * invz,
v1 = dIdy * fy * invz,
v2 = -(v0 * p3d.x + v1 * p3d.y) * invz;
C[0] = v0;
C[1] = v1;
C[2] = v2;
}
inline static
void computeProjectiveMatrix( const Mat& ksi, Mat& Rt )
{
CV_Assert( ksi.size() == Size(1,6) && ksi.type() == CV_64FC1 );
#if defined(HAVE_EIGEN) && EIGEN_WORLD_VERSION == 3 && (!defined _MSC_VER || !defined _M_X64 || _MSC_VER > 1500)
const double* ksi_ptr = reinterpret_cast<const double*>(ksi.ptr(0));
Eigen::Matrix<double,4,4> twist, g;
twist << 0., -ksi_ptr[2], ksi_ptr[1], ksi_ptr[3],
ksi_ptr[2], 0., -ksi_ptr[0], ksi_ptr[4],
-ksi_ptr[1], ksi_ptr[0], 0, ksi_ptr[5],
0., 0., 0., 0.;
g = twist.exp();
eigen2cv(g, Rt);
#else
// for infinitesimal transformation
Rt = Mat::eye(4, 4, CV_64FC1);
Mat R = Rt(Rect(0,0,3,3));
Mat rvec = ksi.rowRange(0,3);
Rodrigues( rvec, R );
Rt.at<double>(0,3) = ksi.at<double>(3);
Rt.at<double>(1,3) = ksi.at<double>(4);
Rt.at<double>(2,3) = ksi.at<double>(5);
#endif
}
static
void cvtDepth2Cloud( const Mat& depth, Mat& cloud, const Mat& cameraMatrix )
{
CV_Assert( cameraMatrix.type() == CV_64FC1 );
const double inv_fx = 1.f/cameraMatrix.at<double>(0,0);
const double inv_fy = 1.f/cameraMatrix.at<double>(1,1);
const double ox = cameraMatrix.at<double>(0,2);
const double oy = cameraMatrix.at<double>(1,2);
cloud.create( depth.size(), CV_32FC3 );
for( int y = 0; y < cloud.rows; y++ )
{
Point3f* cloud_ptr = reinterpret_cast<Point3f*>(cloud.ptr(y));
const float* depth_prt = reinterpret_cast<const float*>(depth.ptr(y));
for( int x = 0; x < cloud.cols; x++ )
{
float z = depth_prt[x];
cloud_ptr[x].x = (float)((x - ox) * z * inv_fx);
cloud_ptr[x].y = (float)((y - oy) * z * inv_fy);
cloud_ptr[x].z = z;
}
}
}
#if SHOW_DEBUG_IMAGES
template<class ImageElemType>
static void warpImage( const Mat& image, const Mat& depth,
const Mat& Rt, const Mat& cameraMatrix, const Mat& distCoeff,
Mat& warpedImage )
{
const Rect rect = Rect(0, 0, image.cols, image.rows);
std::vector<Point2f> points2d;
Mat cloud, transformedCloud;
cvtDepth2Cloud( depth, cloud, cameraMatrix );
perspectiveTransform( cloud, transformedCloud, Rt );
projectPoints( transformedCloud.reshape(3,1), Mat::eye(3,3,CV_64FC1), Mat::zeros(3,1,CV_64FC1), cameraMatrix, distCoeff, points2d );
Mat pointsPositions( points2d );
pointsPositions = pointsPositions.reshape( 2, image.rows );
warpedImage.create( image.size(), image.type() );
warpedImage = Scalar::all(0);
Mat zBuffer( image.size(), CV_32FC1, FLT_MAX );
for( int y = 0; y < image.rows; y++ )
{
for( int x = 0; x < image.cols; x++ )
{
const Point3f p3d = transformedCloud.at<Point3f>(y,x);
const Point p2d = pointsPositions.at<Point2f>(y,x);
if( !cvIsNaN(cloud.at<Point3f>(y,x).z) && cloud.at<Point3f>(y,x).z > 0 &&
rect.contains(p2d) && zBuffer.at<float>(p2d) > p3d.z )
{
warpedImage.at<ImageElemType>(p2d) = image.at<ImageElemType>(y,x);
zBuffer.at<float>(p2d) = p3d.z;
}
}
}
}
#endif
static inline
void set2shorts( int& dst, int short_v1, int short_v2 )
{
unsigned short* ptr = reinterpret_cast<unsigned short*>(&dst);
ptr[0] = static_cast<unsigned short>(short_v1);
ptr[1] = static_cast<unsigned short>(short_v2);
}
static inline
void get2shorts( int src, int& short_v1, int& short_v2 )
{
typedef union { int vint32; unsigned short vuint16[2]; } s32tou16;
const unsigned short* ptr = (reinterpret_cast<s32tou16*>(&src))->vuint16;
short_v1 = ptr[0];
short_v2 = ptr[1];
}
static
int computeCorresp( const Mat& K, const Mat& K_inv, const Mat& Rt,
const Mat& depth0, const Mat& depth1, const Mat& texturedMask1, float maxDepthDiff,
Mat& corresps )
{
CV_Assert( K.type() == CV_64FC1 );
CV_Assert( K_inv.type() == CV_64FC1 );
CV_Assert( Rt.type() == CV_64FC1 );
corresps.create( depth1.size(), CV_32SC1 );
Mat R = Rt(Rect(0,0,3,3)).clone();
Mat KRK_inv = K * R * K_inv;
const double * KRK_inv_ptr = reinterpret_cast<const double *>(KRK_inv.ptr());
Mat Kt = Rt(Rect(3,0,1,3)).clone();
Kt = K * Kt;
const double * Kt_ptr = reinterpret_cast<const double *>(Kt.ptr());
Rect r(0, 0, depth1.cols, depth1.rows);
corresps = Scalar(-1);
int correspCount = 0;
for( int v1 = 0; v1 < depth1.rows; v1++ )
{
for( int u1 = 0; u1 < depth1.cols; u1++ )
{
float d1 = depth1.at<float>(v1,u1);
if( !cvIsNaN(d1) && texturedMask1.at<uchar>(v1,u1) )
{
float transformed_d1 = (float)(d1 * (KRK_inv_ptr[6] * u1 + KRK_inv_ptr[7] * v1 + KRK_inv_ptr[8]) + Kt_ptr[2]);
int u0 = cvRound((d1 * (KRK_inv_ptr[0] * u1 + KRK_inv_ptr[1] * v1 + KRK_inv_ptr[2]) + Kt_ptr[0]) / transformed_d1);
int v0 = cvRound((d1 * (KRK_inv_ptr[3] * u1 + KRK_inv_ptr[4] * v1 + KRK_inv_ptr[5]) + Kt_ptr[1]) / transformed_d1);
if( r.contains(Point(u0,v0)) )
{
float d0 = depth0.at<float>(v0,u0);
if( !cvIsNaN(d0) && std::abs(transformed_d1 - d0) <= maxDepthDiff )
{
int c = corresps.at<int>(v0,u0);
if( c != -1 )
{
int exist_u1, exist_v1;
get2shorts( c, exist_u1, exist_v1);
float exist_d1 = (float)(depth1.at<float>(exist_v1,exist_u1) * (KRK_inv_ptr[6] * exist_u1 + KRK_inv_ptr[7] * exist_v1 + KRK_inv_ptr[8]) + Kt_ptr[2]);
if( transformed_d1 > exist_d1 )
continue;
}
else
correspCount++;
set2shorts( corresps.at<int>(v0,u0), u1, v1 );
}
}
}
}
}
return correspCount;
}
static inline
void preprocessDepth( Mat depth0, Mat depth1,
const Mat& validMask0, const Mat& validMask1,
float minDepth, float maxDepth )
{
CV_DbgAssert( depth0.size() == depth1.size() );
for( int y = 0; y < depth0.rows; y++ )
{
for( int x = 0; x < depth0.cols; x++ )
{
float& d0 = depth0.at<float>(y,x);
if( !cvIsNaN(d0) && (d0 > maxDepth || d0 < minDepth || d0 <= 0 || (!validMask0.empty() && !validMask0.at<uchar>(y,x))) )
d0 = std::numeric_limits<float>::quiet_NaN();
float& d1 = depth1.at<float>(y,x);
if( !cvIsNaN(d1) && (d1 > maxDepth || d1 < minDepth || d1 <= 0 || (!validMask1.empty() && !validMask1.at<uchar>(y,x))) )
d1 = std::numeric_limits<float>::quiet_NaN();
}
}
}
static
void buildPyramids( const Mat& image0, const Mat& image1,
const Mat& depth0, const Mat& depth1,
const Mat& cameraMatrix, int sobelSize, double sobelScale,
const std::vector<float>& minGradMagnitudes,
std::vector<Mat>& pyramidImage0, std::vector<Mat>& pyramidDepth0,
std::vector<Mat>& pyramidImage1, std::vector<Mat>& pyramidDepth1,
std::vector<Mat>& pyramid_dI_dx1, std::vector<Mat>& pyramid_dI_dy1,
std::vector<Mat>& pyramidTexturedMask1, std::vector<Mat>& pyramidCameraMatrix )
{
const int pyramidMaxLevel = (int)minGradMagnitudes.size() - 1;
buildPyramid( image0, pyramidImage0, pyramidMaxLevel );
buildPyramid( image1, pyramidImage1, pyramidMaxLevel );
pyramid_dI_dx1.resize( pyramidImage1.size() );
pyramid_dI_dy1.resize( pyramidImage1.size() );
pyramidTexturedMask1.resize( pyramidImage1.size() );
pyramidCameraMatrix.reserve( pyramidImage1.size() );
Mat cameraMatrix_dbl;
cameraMatrix.convertTo( cameraMatrix_dbl, CV_64FC1 );
for( size_t i = 0; i < pyramidImage1.size(); i++ )
{
Sobel( pyramidImage1[i], pyramid_dI_dx1[i], CV_16S, 1, 0, sobelSize );
Sobel( pyramidImage1[i], pyramid_dI_dy1[i], CV_16S, 0, 1, sobelSize );
const Mat& dx = pyramid_dI_dx1[i];
const Mat& dy = pyramid_dI_dy1[i];
Mat texturedMask( dx.size(), CV_8UC1, Scalar(0) );
const float minScalesGradMagnitude2 = (float)((minGradMagnitudes[i] * minGradMagnitudes[i]) / (sobelScale * sobelScale));
for( int y = 0; y < dx.rows; y++ )
{
for( int x = 0; x < dx.cols; x++ )
{
float m2 = (float)(dx.at<short>(y,x)*dx.at<short>(y,x) + dy.at<short>(y,x)*dy.at<short>(y,x));
if( m2 >= minScalesGradMagnitude2 )
texturedMask.at<uchar>(y,x) = 255;
}
}
pyramidTexturedMask1[i] = texturedMask;
Mat levelCameraMatrix = i == 0 ? cameraMatrix_dbl : 0.5f * pyramidCameraMatrix[i-1];
levelCameraMatrix.at<double>(2,2) = 1.;
pyramidCameraMatrix.push_back( levelCameraMatrix );
}
buildPyramid( depth0, pyramidDepth0, pyramidMaxLevel );
buildPyramid( depth1, pyramidDepth1, pyramidMaxLevel );
}
static
bool solveSystem( const Mat& C, const Mat& dI_dt, double detThreshold, Mat& ksi )
{
#if defined(HAVE_EIGEN) && EIGEN_WORLD_VERSION == 3
Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> eC, eCt, edI_dt;
cv2eigen(C, eC);
cv2eigen(dI_dt, edI_dt);
eCt = eC.transpose();
Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> A, B, eksi;
A = eCt * eC;
double det = A.determinant();
if( fabs (det) < detThreshold || cvIsNaN(det) || cvIsInf(det) )
return false;
B = -eCt * edI_dt;
eksi = A.ldlt().solve(B);
eigen2cv( eksi, ksi );
#else
Mat A = C.t() * C;
double det = cv::determinant(A);
if( fabs (det) < detThreshold || cvIsNaN(det) || cvIsInf(det) )
return false;
Mat B = -C.t() * dI_dt;
cv::solve( A, B, ksi, DECOMP_CHOLESKY );
#endif
return true;
}
typedef void (*ComputeCFuncPtr)( double* C, double dIdx, double dIdy, const Point3f& p3d, double fx, double fy );
static
bool computeKsi( int transformType,
const Mat& image0, const Mat& cloud0,
const Mat& image1, const Mat& dI_dx1, const Mat& dI_dy1,
const Mat& corresps, int correspsCount,
double fx, double fy, double sobelScale, double determinantThreshold,
Mat& ksi )
{
int Cwidth = -1;
ComputeCFuncPtr computeCFuncPtr = 0;
if( transformType == RIGID_BODY_MOTION )
{
Cwidth = 6;
computeCFuncPtr = computeC_RigidBodyMotion;
}
else if( transformType == ROTATION )
{
Cwidth = 3;
computeCFuncPtr = computeC_Rotation;
}
else if( transformType == TRANSLATION )
{
Cwidth = 3;
computeCFuncPtr = computeC_Translation;
}
else
CV_Error(Error::StsBadFlag, "Unsupported value of transformation type flag.");
Mat C( correspsCount, Cwidth, CV_64FC1 );
Mat dI_dt( correspsCount, 1, CV_64FC1 );
double sigma = 0;
int pointCount = 0;
for( int v0 = 0; v0 < corresps.rows; v0++ )
{
for( int u0 = 0; u0 < corresps.cols; u0++ )
{
if( corresps.at<int>(v0,u0) != -1 )
{
int u1, v1;
get2shorts( corresps.at<int>(v0,u0), u1, v1 );
double diff = static_cast<double>(image1.at<uchar>(v1,u1)) -
static_cast<double>(image0.at<uchar>(v0,u0));
sigma += diff * diff;
pointCount++;
}
}
}
sigma = std::sqrt(sigma/pointCount);
pointCount = 0;
for( int v0 = 0; v0 < corresps.rows; v0++ )
{
for( int u0 = 0; u0 < corresps.cols; u0++ )
{
if( corresps.at<int>(v0,u0) != -1 )
{
int u1, v1;
get2shorts( corresps.at<int>(v0,u0), u1, v1 );
double diff = static_cast<double>(image1.at<uchar>(v1,u1)) -
static_cast<double>(image0.at<uchar>(v0,u0));
double w = sigma + std::abs(diff);
w = w > DBL_EPSILON ? 1./w : 1.;
(*computeCFuncPtr)( (double*)C.ptr(pointCount),
w * sobelScale * dI_dx1.at<short int>(v1,u1),
w * sobelScale * dI_dy1.at<short int>(v1,u1),
cloud0.at<Point3f>(v0,u0), fx, fy);
dI_dt.at<double>(pointCount) = w * diff;
pointCount++;
}
}
}
Mat sln;
bool solutionExist = solveSystem( C, dI_dt, determinantThreshold, sln );
if( solutionExist )
{
ksi.create(6,1,CV_64FC1);
ksi = Scalar(0);
Mat subksi;
if( transformType == RIGID_BODY_MOTION )
{
subksi = ksi;
}
else if( transformType == ROTATION )
{
subksi = ksi.rowRange(0,3);
}
else if( transformType == TRANSLATION )
{
subksi = ksi.rowRange(3,6);
}
sln.copyTo( subksi );
}
return solutionExist;
}
bool cv::RGBDOdometry( cv::Mat& Rt, const Mat& initRt,
const cv::Mat& image0, const cv::Mat& _depth0, const cv::Mat& validMask0,
const cv::Mat& image1, const cv::Mat& _depth1, const cv::Mat& validMask1,
const cv::Mat& cameraMatrix, float minDepth, float maxDepth, float maxDepthDiff,
const std::vector<int>& iterCounts, const std::vector<float>& minGradientMagnitudes,
int transformType )
{
const int sobelSize = 3;
const double sobelScale = 1./8;
Mat depth0 = _depth0.clone(),
depth1 = _depth1.clone();
// check RGB-D input data
CV_Assert( !image0.empty() );
CV_Assert( image0.type() == CV_8UC1 );
CV_Assert( depth0.type() == CV_32FC1 && depth0.size() == image0.size() );
CV_Assert( image1.size() == image0.size() );
CV_Assert( image1.type() == CV_8UC1 );
CV_Assert( depth1.type() == CV_32FC1 && depth1.size() == image0.size() );
// check masks
CV_Assert( validMask0.empty() || (validMask0.type() == CV_8UC1 && validMask0.size() == image0.size()) );
CV_Assert( validMask1.empty() || (validMask1.type() == CV_8UC1 && validMask1.size() == image0.size()) );
// check camera params
CV_Assert( cameraMatrix.type() == CV_32FC1 && cameraMatrix.size() == Size(3,3) );
// other checks
CV_Assert( iterCounts.empty() || minGradientMagnitudes.empty() ||
minGradientMagnitudes.size() == iterCounts.size() );
CV_Assert( initRt.empty() || (initRt.type()==CV_64FC1 && initRt.size()==Size(4,4) ) );
std::vector<int> defaultIterCounts;
std::vector<float> defaultMinGradMagnitudes;
std::vector<int> const* iterCountsPtr = &iterCounts;
std::vector<float> const* minGradientMagnitudesPtr = &minGradientMagnitudes;
if( iterCounts.empty() || minGradientMagnitudes.empty() )
{
defaultIterCounts.resize(4);
defaultIterCounts[0] = 7;
defaultIterCounts[1] = 7;
defaultIterCounts[2] = 7;
defaultIterCounts[3] = 10;
defaultMinGradMagnitudes.resize(4);
defaultMinGradMagnitudes[0] = 12;
defaultMinGradMagnitudes[1] = 5;
defaultMinGradMagnitudes[2] = 3;
defaultMinGradMagnitudes[3] = 1;
iterCountsPtr = &defaultIterCounts;
minGradientMagnitudesPtr = &defaultMinGradMagnitudes;
}
preprocessDepth( depth0, depth1, validMask0, validMask1, minDepth, maxDepth );
std::vector<Mat> pyramidImage0, pyramidDepth0,
pyramidImage1, pyramidDepth1, pyramid_dI_dx1, pyramid_dI_dy1, pyramidTexturedMask1,
pyramidCameraMatrix;
buildPyramids( image0, image1, depth0, depth1, cameraMatrix, sobelSize, sobelScale, *minGradientMagnitudesPtr,
pyramidImage0, pyramidDepth0, pyramidImage1, pyramidDepth1,
pyramid_dI_dx1, pyramid_dI_dy1, pyramidTexturedMask1, pyramidCameraMatrix );
Mat resultRt = initRt.empty() ? Mat::eye(4,4,CV_64FC1) : initRt.clone();
Mat currRt, ksi;
for( int level = (int)iterCountsPtr->size() - 1; level >= 0; level-- )
{
const Mat& levelCameraMatrix = pyramidCameraMatrix[level];
const Mat& levelImage0 = pyramidImage0[level];
const Mat& levelDepth0 = pyramidDepth0[level];
Mat levelCloud0;
cvtDepth2Cloud( pyramidDepth0[level], levelCloud0, levelCameraMatrix );
const Mat& levelImage1 = pyramidImage1[level];
const Mat& levelDepth1 = pyramidDepth1[level];
const Mat& level_dI_dx1 = pyramid_dI_dx1[level];
const Mat& level_dI_dy1 = pyramid_dI_dy1[level];
CV_Assert( level_dI_dx1.type() == CV_16S );
CV_Assert( level_dI_dy1.type() == CV_16S );
const double fx = levelCameraMatrix.at<double>(0,0);
const double fy = levelCameraMatrix.at<double>(1,1);
const double determinantThreshold = 1e-6;
Mat corresps( levelImage0.size(), levelImage0.type() );
// Run transformation search on current level iteratively.
for( int iter = 0; iter < (*iterCountsPtr)[level]; iter ++ )
{
int correspsCount = computeCorresp( levelCameraMatrix, levelCameraMatrix.inv(), resultRt.inv(DECOMP_SVD),
levelDepth0, levelDepth1, pyramidTexturedMask1[level], maxDepthDiff,
corresps );
if( correspsCount == 0 )
break;
bool solutionExist = computeKsi( transformType,
levelImage0, levelCloud0,
levelImage1, level_dI_dx1, level_dI_dy1,
corresps, correspsCount,
fx, fy, sobelScale, determinantThreshold,
ksi );
if( !solutionExist )
break;
computeProjectiveMatrix( ksi, currRt );
resultRt = currRt * resultRt;
#if SHOW_DEBUG_IMAGES
std::cout << "currRt " << currRt << std::endl;
Mat warpedImage0;
const Mat distCoeff(1,5,CV_32FC1,Scalar(0));
warpImage<uchar>( levelImage0, levelDepth0, resultRt, levelCameraMatrix, distCoeff, warpedImage0 );
imshow( "im0", levelImage0 );
imshow( "wim0", warpedImage0 );
imshow( "im1", levelImage1 );
waitKey();
#endif
}
}
Rt = resultRt;
return !Rt.empty();
}

259
modules/contrib/src/selfsimilarity.cpp
View File

@@ -1,259 +0,0 @@
// This is based on Rainer Lienhart contribution. Below is the original copyright:
//
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// University of Augsburg License Agreement
// For Open Source MultiMedia Computing (MMC) Library
//
// Copyright (C) 2007, University of Augsburg, Germany, all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of University of Augsburg, Germany may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the University of Augsburg, Germany or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
// Author: Rainer Lienhart
// email: Rainer.Lienhart@informatik.uni-augsburg.de
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
// Please cite the following two papers:
// 1. Shechtman, E., Irani, M.:
// Matching local self-similarities across images and videos.
// CVPR, (2007)
// 2. Eva Horster, Thomas Greif, Rainer Lienhart, Malcolm Slaney.
// Comparing Local Feature Descriptors in pLSA-Based Image Models.
// 30th Annual Symposium of the German Association for
// Pattern Recognition (DAGM) 2008, Munich, Germany, June 2008.
#include "precomp.hpp"
namespace cv
{
SelfSimDescriptor::SelfSimDescriptor()
{
smallSize = DEFAULT_SMALL_SIZE;
largeSize = DEFAULT_LARGE_SIZE;
numberOfAngles = DEFAULT_NUM_ANGLES;
startDistanceBucket = DEFAULT_START_DISTANCE_BUCKET;
numberOfDistanceBuckets = DEFAULT_NUM_DISTANCE_BUCKETS;
}
SelfSimDescriptor::SelfSimDescriptor(int _ssize, int _lsize,
int _startDistanceBucket,
int _numberOfDistanceBuckets, int _numberOfAngles)
{
smallSize = _ssize;
largeSize = _lsize;
startDistanceBucket = _startDistanceBucket;
numberOfDistanceBuckets = _numberOfDistanceBuckets;
numberOfAngles = _numberOfAngles;
}
SelfSimDescriptor::SelfSimDescriptor(const SelfSimDescriptor& ss)
{
smallSize = ss.smallSize;
largeSize = ss.largeSize;
startDistanceBucket = ss.startDistanceBucket;
numberOfDistanceBuckets = ss.numberOfDistanceBuckets;
numberOfAngles = ss.numberOfAngles;
}
SelfSimDescriptor::~SelfSimDescriptor()
{
}
SelfSimDescriptor& SelfSimDescriptor::operator = (const SelfSimDescriptor& ss)
{
if( this != &ss )
{
smallSize = ss.smallSize;
largeSize = ss.largeSize;
startDistanceBucket = ss.startDistanceBucket;
numberOfDistanceBuckets = ss.numberOfDistanceBuckets;
numberOfAngles = ss.numberOfAngles;
}
return *this;
}
size_t SelfSimDescriptor::getDescriptorSize() const
{
return numberOfAngles*(numberOfDistanceBuckets - startDistanceBucket);
}
Size SelfSimDescriptor::getGridSize( Size imgSize, Size winStride ) const
{
winStride.width = std::max(winStride.width, 1);
winStride.height = std::max(winStride.height, 1);
int border = largeSize/2 + smallSize/2;
return Size(std::max(imgSize.width - border*2 + winStride.width - 1, 0)/winStride.width,
std::max(imgSize.height - border*2 + winStride.height - 1, 0)/winStride.height);
}
// TODO: optimized with SSE2
void SelfSimDescriptor::SSD(const Mat& img, Point pt, Mat& ssd) const
{
int x, y, dx, dy, r0 = largeSize/2, r1 = smallSize/2;
int step = (int)img.step;
for( y = -r0; y <= r0; y++ )
{
float* sptr = ssd.ptr<float>(y+r0) + r0;
for( x = -r0; x <= r0; x++ )
{
int sum = 0;
const uchar* src0 = img.ptr<uchar>(y + pt.y - r1) + x + pt.x;
const uchar* src1 = img.ptr<uchar>(pt.y - r1) + pt.x;
for( dy = -r1; dy <= r1; dy++, src0 += step, src1 += step )
for( dx = -r1; dx <= r1; dx++ )
{
int t = src0[dx] - src1[dx];
sum += t*t;
}
sptr[x] = (float)sum;
}
}
}
void SelfSimDescriptor::compute(const Mat& img, std::vector<float>& descriptors, Size winStride,
const std::vector<Point>& locations) const
{
CV_Assert( img.depth() == CV_8U );
winStride.width = std::max(winStride.width, 1);
winStride.height = std::max(winStride.height, 1);
Size gridSize = getGridSize(img.size(), winStride);
int i, nwindows = locations.empty() ? gridSize.width*gridSize.height : (int)locations.size();
int border = largeSize/2 + smallSize/2;
int fsize = (int)getDescriptorSize();
std::vector<float> tempFeature(fsize+1);
descriptors.resize(fsize*nwindows + 1);
Mat ssd(largeSize, largeSize, CV_32F), mappingMask;
computeLogPolarMapping(mappingMask);
#if 0 //def _OPENMP
int nthreads = cvGetNumThreads();
#pragma omp parallel for num_threads(nthreads)
#endif
for( i = 0; i < nwindows; i++ )
{
Point pt;
float* feature0 = &descriptors[fsize*i];
float* feature = &tempFeature[0];
int x, y, j;
if( !locations.empty() )
{
pt = locations[i];
if( pt.x < border || pt.x >= img.cols - border ||
pt.y < border || pt.y >= img.rows - border )
{
for( j = 0; j < fsize; j++ )
feature0[j] = 0.f;
continue;
}
}
else
pt = Point((i % gridSize.width)*winStride.width + border,
(i / gridSize.width)*winStride.height + border);
SSD(img, pt, ssd);
// Determine in the local neighborhood the largest difference and use for normalization
float var_noise = 1000.f;
for( y = -1; y <= 1 ; y++ )
for( x = -1 ; x <= 1 ; x++ )
var_noise = std::max(var_noise, ssd.at<float>(largeSize/2+y, largeSize/2+x));
for( j = 0; j <= fsize; j++ )
feature[j] = FLT_MAX;
// Derive feature vector before exp(-x) computation
// Idea: for all x,a >= 0, a=const. we have:
// max [ exp( -x / a) ] = exp ( -min(x) / a )
// Thus, determine min(ssd) and store in feature[...]
for( y = 0; y < ssd.rows; y++ )
{
const schar *mappingMaskPtr = mappingMask.ptr<schar>(y);
const float *ssdPtr = ssd.ptr<float>(y);
for( x = 0 ; x < ssd.cols; x++ )
{
int index = mappingMaskPtr[x];
feature[index] = std::min(feature[index], ssdPtr[x]);
}
}
var_noise = -1.f/var_noise;
for( j = 0; j < fsize; j++ )
feature0[j] = feature[j]*var_noise;
Mat _f(1, fsize, CV_32F, feature0);
cv::exp(_f, _f);
}
}
void SelfSimDescriptor::computeLogPolarMapping(Mat& mappingMask) const
{
mappingMask.create(largeSize, largeSize, CV_8S);
// What we want is
// log_m (radius) = numberOfDistanceBuckets
// <==> log_10 (radius) / log_10 (m) = numberOfDistanceBuckets
// <==> log_10 (radius) / numberOfDistanceBuckets = log_10 (m)
// <==> m = 10 ^ log_10(m) = 10 ^ [log_10 (radius) / numberOfDistanceBuckets]
//
int radius = largeSize/2, angleBucketSize = 360 / numberOfAngles;
int fsize = (int)getDescriptorSize();
double inv_log10m = (double)numberOfDistanceBuckets/log10((double)radius);
for (int y=-radius ; y<=radius ; y++)
{
schar* mrow = mappingMask.ptr<schar>(y+radius);
for (int x=-radius ; x<=radius ; x++)
{
int index = fsize;
float dist = (float)std::sqrt((float)x*x + (float)y*y);
int distNo = dist > 0 ? cvRound(log10(dist)*inv_log10m) : 0;
if( startDistanceBucket <= distNo && distNo < numberOfDistanceBuckets )
{
float angle = std::atan2( (float)y, (float)x ) / (float)CV_PI * 180.0f;
if (angle < 0) angle += 360.0f;
int angleInt = (cvRound(angle) + angleBucketSize/2) % 360;
int angleIndex = angleInt / angleBucketSize;
index = (distNo-startDistanceBucket)*numberOfAngles + angleIndex;
}
mrow[x + radius] = saturate_cast<schar>(index);
}
}
}
}

1212
modules/contrib/src/spinimages.cpp
View File

File diff suppressed because it is too large Load Diff

411
modules/contrib/src/stereovar.cpp
View File

@@ -1,411 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
/*
This is a modification of the variational stereo correspondence algorithm, described in:
S. Kosov, T. Thormaehlen, H.-P. Seidel "Accurate Real-Time Disparity Estimation with Variational Methods"
Proceedings of the 5th International Symposium on Visual Computing, Vegas, USA
This code is written by Sergey G. Kosov for "Visir PX" application as part of Project X (www.project-10.de)
*/
#include "precomp.hpp"
#include <limits.h>
namespace cv
{
StereoVar::StereoVar() : levels(3), pyrScale(0.5), nIt(5), minDisp(0), maxDisp(16), poly_n(3), poly_sigma(0), fi(25.0f), lambda(0.03f), penalization(PENALIZATION_TICHONOV), cycle(CYCLE_V), flags(USE_SMART_ID | USE_AUTO_PARAMS)
{
}
StereoVar::StereoVar(int _levels, double _pyrScale, int _nIt, int _minDisp, int _maxDisp, int _poly_n, double _poly_sigma, float _fi, float _lambda, int _penalization, int _cycle, int _flags) : levels(_levels), pyrScale(_pyrScale), nIt(_nIt), minDisp(_minDisp), maxDisp(_maxDisp), poly_n(_poly_n), poly_sigma(_poly_sigma), fi(_fi), lambda(_lambda), penalization(_penalization), cycle(_cycle), flags(_flags)
{ // No Parameters check, since they are all public
}
StereoVar::~StereoVar()
{
}
static Mat diffX(Mat &src)
{
int cols = src.cols - 1;
Mat dst(src.size(), src.type());
for(int y = 0; y < src.rows; y++){
const float* pSrc = src.ptr<float>(y);
float* pDst = dst.ptr<float>(y);
int x = 0;
#if CV_SSE2
for (x = 0; x <= cols - 8; x += 8) {
__m128 a0 = _mm_loadu_ps(pSrc + x);
__m128 b0 = _mm_loadu_ps(pSrc + x + 1);
__m128 a1 = _mm_loadu_ps(pSrc + x + 4);
__m128 b1 = _mm_loadu_ps(pSrc + x + 5);
b0 = _mm_sub_ps(b0, a0);
b1 = _mm_sub_ps(b1, a1);
_mm_storeu_ps(pDst + x, b0);
_mm_storeu_ps(pDst + x + 4, b1);
}
#endif
for( ; x < cols; x++) pDst[x] = pSrc[x+1] - pSrc[x];
pDst[cols] = 0.f;
}
return dst;
}
static Mat getGradient(Mat &src)
{
register int x, y;
Mat dst(src.size(), src.type());
dst.setTo(0);
for (y = 0; y < src.rows - 1; y++) {
float *pSrc = src.ptr<float>(y);
float *pSrcF = src.ptr<float>(y + 1);
float *pDst = dst.ptr<float>(y);
for (x = 0; x < src.cols - 1; x++)
pDst[x] = fabs(pSrc[x + 1] - pSrc[x]) + fabs(pSrcF[x] - pSrc[x]);
}
return dst;
}
static Mat getG_c(Mat &src, float l)
{
Mat dst(src.size(), src.type());
for (register int y = 0; y < src.rows; y++) {
float *pSrc = src.ptr<float>(y);
float *pDst = dst.ptr<float>(y);
for (register int x = 0; x < src.cols; x++)
pDst[x] = 0.5f*l / sqrtf(l*l + pSrc[x]*pSrc[x]);
}
return dst;
}
static Mat getG_p(Mat &src, float l)
{
Mat dst(src.size(), src.type());
for (register int y = 0; y < src.rows; y++) {
float *pSrc = src.ptr<float>(y);
float *pDst = dst.ptr<float>(y);
for (register int x = 0; x < src.cols; x++)
pDst[x] = 0.5f*l*l / (l*l + pSrc[x]*pSrc[x]);
}
return dst;
}
void StereoVar::VariationalSolver(Mat &I1, Mat &I2, Mat &I2x, Mat &u, int level)
{
register int n, x, y;
float gl = 1, gr = 1, gu = 1, gd = 1, gc = 4;
Mat g_c, g_p;
Mat U;
u.copyTo(U);
int N = nIt;
float l = lambda;
float Fi = fi;
if (flags & USE_SMART_ID) {
double scale = std::pow(pyrScale, (double) level) * (1 + pyrScale);
N = (int) (N / scale);
}
double scale = std::pow(pyrScale, (double) level);
Fi /= (float) scale;
l *= (float) scale;
int width = u.cols - 1;
int height = u.rows - 1;
for (n = 0; n < N; n++) {
if (penalization != PENALIZATION_TICHONOV) {
Mat gradient = getGradient(U);
switch (penalization) {
case PENALIZATION_CHARBONNIER: g_c = getG_c(gradient, l); break;
case PENALIZATION_PERONA_MALIK: g_p = getG_p(gradient, l); break;
}
gradient.release();
}
for (y = 1 ; y < height; y++) {
float *pU = U.ptr<float>(y);
float *pUu = U.ptr<float>(y + 1);
float *pUd = U.ptr<float>(y - 1);
float *pu = u.ptr<float>(y);
float *pI1 = I1.ptr<float>(y);
float *pI2 = I2.ptr<float>(y);
float *pI2x = I2x.ptr<float>(y);
float *pG_c = NULL, *pG_cu = NULL, *pG_cd = NULL;
float *pG_p = NULL, *pG_pu = NULL, *pG_pd = NULL;
switch (penalization) {
case PENALIZATION_CHARBONNIER:
pG_c = g_c.ptr<float>(y);
pG_cu = g_c.ptr<float>(y + 1);
pG_cd = g_c.ptr<float>(y - 1);
break;
case PENALIZATION_PERONA_MALIK:
pG_p = g_p.ptr<float>(y);
pG_pu = g_p.ptr<float>(y + 1);
pG_pd = g_p.ptr<float>(y - 1);
break;
}
for (x = 1; x < width; x++) {
switch (penalization) {
case PENALIZATION_CHARBONNIER:
gc = pG_c[x];
gl = gc + pG_c[x - 1];
gr = gc + pG_c[x + 1];
gu = gc + pG_cu[x];
gd = gc + pG_cd[x];
gc = gl + gr + gu + gd;
break;
case PENALIZATION_PERONA_MALIK:
gc = pG_p[x];
gl = gc + pG_p[x - 1];
gr = gc + pG_p[x + 1];
gu = gc + pG_pu[x];
gd = gc + pG_pd[x];
gc = gl + gr + gu + gd;
break;
}
float _fi = Fi;
if (maxDisp > minDisp) {
if (pU[x] > maxDisp * scale) {_fi *= 1000; pU[x] = static_cast<float>(maxDisp * scale);}
if (pU[x] < minDisp * scale) {_fi *= 1000; pU[x] = static_cast<float>(minDisp * scale);}
}
int A = static_cast<int>(pU[x]);
int neg = 0; if (pU[x] <= 0) neg = -1;
if (x + A > width)
pu[x] = pU[width - A];
else if (x + A + neg < 0)
pu[x] = pU[- A + 2];
else {
pu[x] = A + (pI2x[x + A + neg] * (pI1[x] - pI2[x + A])
+ _fi * (gr * pU[x + 1] + gl * pU[x - 1] + gu * pUu[x] + gd * pUd[x] - gc * A))
/ (pI2x[x + A + neg] * pI2x[x + A + neg] + gc * _fi) ;
}
}// x
pu[0] = pu[1];
pu[width] = pu[width - 1];
}// y
for (x = 0; x <= width; x++) {
u.at<float>(0, x) = u.at<float>(1, x);
u.at<float>(height, x) = u.at<float>(height - 1, x);
}
u.copyTo(U);
if (!g_c.empty()) g_c.release();
if (!g_p.empty()) g_p.release();
}//n
}
void StereoVar::VCycle_MyFAS(Mat &I1, Mat &I2, Mat &I2x, Mat &_u, int level)
{
Size imgSize = _u.size();
Size frmSize = Size((int) (imgSize.width * pyrScale + 0.5), (int) (imgSize.height * pyrScale + 0.5));
Mat I1_h, I2_h, I2x_h, u_h, U, U_h;
//PRE relaxation
VariationalSolver(I1, I2, I2x, _u, level);
if (level >= levels - 1) return;
level ++;
//scaling DOWN
resize(I1, I1_h, frmSize, 0, 0, INTER_AREA);
resize(I2, I2_h, frmSize, 0, 0, INTER_AREA);
resize(_u, u_h, frmSize, 0, 0, INTER_AREA);
u_h.convertTo(u_h, u_h.type(), pyrScale);
I2x_h = diffX(I2_h);
//Next level
U_h = u_h.clone();
VCycle_MyFAS(I1_h, I2_h, I2x_h, U_h, level);
subtract(U_h, u_h, U_h);
U_h.convertTo(U_h, U_h.type(), 1.0 / pyrScale);
//scaling UP
resize(U_h, U, imgSize);
//correcting the solution
add(_u, U, _u);
//POST relaxation
VariationalSolver(I1, I2, I2x, _u, level - 1);
if (flags & USE_MEDIAN_FILTERING) medianBlur(_u, _u, 3);
I1_h.release();
I2_h.release();
I2x_h.release();
u_h.release();
U.release();
U_h.release();
}
void StereoVar::FMG(Mat &I1, Mat &I2, Mat &I2x, Mat &u, int level)
{
double scale = std::pow(pyrScale, (double) level);
Size frmSize = Size((int) (u.cols * scale + 0.5), (int) (u.rows * scale + 0.5));
Mat I1_h, I2_h, I2x_h, u_h;
//scaling DOWN
resize(I1, I1_h, frmSize, 0, 0, INTER_AREA);
resize(I2, I2_h, frmSize, 0, 0, INTER_AREA);
resize(u, u_h, frmSize, 0, 0, INTER_AREA);
u_h.convertTo(u_h, u_h.type(), scale);
I2x_h = diffX(I2_h);
switch (cycle) {
case CYCLE_O:
VariationalSolver(I1_h, I2_h, I2x_h, u_h, level);
break;
case CYCLE_V:
VCycle_MyFAS(I1_h, I2_h, I2x_h, u_h, level);
break;
}
u_h.convertTo(u_h, u_h.type(), 1.0 / scale);
//scaling UP
resize(u_h, u, u.size(), 0, 0, INTER_CUBIC);
I1_h.release();
I2_h.release();
I2x_h.release();
u_h.release();
level--;
if ((flags & USE_AUTO_PARAMS) && (level < levels / 3)) {
penalization = PENALIZATION_PERONA_MALIK;
fi *= 100;
flags -= USE_AUTO_PARAMS;
autoParams();
}
if (flags & USE_MEDIAN_FILTERING) medianBlur(u, u, 3);
if (level >= 0) FMG(I1, I2, I2x, u, level);
}
void StereoVar::autoParams()
{
int maxD = MAX(labs(maxDisp), labs(minDisp));
if (!maxD) pyrScale = 0.85;
else if (maxD < 8) pyrScale = 0.5;
else if (maxD < 64) pyrScale = 0.5 + static_cast<double>(maxD - 8) * 0.00625;
else pyrScale = 0.85;
if (maxD) {
levels = 0;
while ( std::pow(pyrScale, levels) * maxD > 1.5) levels ++;
levels++;
}
switch(penalization) {
case PENALIZATION_TICHONOV: cycle = CYCLE_V; break;
case PENALIZATION_CHARBONNIER: cycle = CYCLE_O; break;
case PENALIZATION_PERONA_MALIK: cycle = CYCLE_O; break;
}
}
void StereoVar::operator ()( const Mat& left, const Mat& right, Mat& disp )
{
CV_Assert(left.size() == right.size() && left.type() == right.type());
Size imgSize = left.size();
int MaxD = MAX(labs(minDisp), labs(maxDisp));
int SignD = 1; if (MIN(minDisp, maxDisp) < 0) SignD = -1;
if (minDisp >= maxDisp) {MaxD = 256; SignD = 1;}
Mat u;
if ((flags & USE_INITIAL_DISPARITY) && (!disp.empty())) {
CV_Assert(disp.size() == left.size() && disp.type() == CV_8UC1);
disp.convertTo(u, CV_32FC1, static_cast<double>(SignD * MaxD) / 256);
} else {
u.create(imgSize, CV_32FC1);
u.setTo(0);
}
// Preprocessing
Mat leftgray, rightgray;
if (left.type() != CV_8UC1) {
cvtColor(left, leftgray, COLOR_BGR2GRAY);
cvtColor(right, rightgray, COLOR_BGR2GRAY);
} else {
left.copyTo(leftgray);
right.copyTo(rightgray);
}
if (flags & USE_EQUALIZE_HIST) {
equalizeHist(leftgray, leftgray);
equalizeHist(rightgray, rightgray);
}
if (poly_sigma > 0.0001) {
GaussianBlur(leftgray, leftgray, Size(poly_n, poly_n), poly_sigma);
GaussianBlur(rightgray, rightgray, Size(poly_n, poly_n), poly_sigma);
}
if (flags & USE_AUTO_PARAMS) {
penalization = PENALIZATION_TICHONOV;
autoParams();
}
Mat I1, I2;
leftgray.convertTo(I1, CV_32FC1);
rightgray.convertTo(I2, CV_32FC1);
leftgray.release();
rightgray.release();
Mat I2x = diffX(I2);
FMG(I1, I2, I2x, u, levels - 1);
I1.release();
I2.release();
I2x.release();
disp.create( left.size(), CV_8UC1 );
u = abs(u);
u.convertTo(disp, disp.type(), 256 / MaxD, 0);
u.release();
}
} // namespace

2
modules/core/doc/basic_structures.rst
View File

@@ -2981,7 +2981,7 @@ The class provides the following features for all derived classes:
* so called "virtual constructor". That is, each Algorithm derivative is registered at program start and you can get the list of registered algorithms and create instance of a particular algorithm by its name (see ``Algorithm::create``). If you plan to add your own algorithms, it is good practice to add a unique prefix to your algorithms to distinguish them from other algorithms.
* setting/retrieving algorithm parameters by name. If you used video capturing functionality from OpenCV highgui module, you are probably familar with ``cvSetCaptureProperty()``, ``cvGetCaptureProperty()``, ``VideoCapture::set()`` and ``VideoCapture::get()``. ``Algorithm`` provides similar method where instead of integer id's you specify the parameter names as text strings. See ``Algorithm::set`` and ``Algorithm::get`` for details.
* setting/retrieving algorithm parameters by name. If you used video capturing functionality from OpenCV videoio module, you are probably familar with ``cvSetCaptureProperty()``, ``cvGetCaptureProperty()``, ``VideoCapture::set()`` and ``VideoCapture::get()``. ``Algorithm`` provides similar method where instead of integer id's you specify the parameter names as text strings. See ``Algorithm::set`` and ``Algorithm::get`` for details.
* reading and writing parameters from/to XML or YAML files. Every Algorithm derivative can store all its parameters and then read them back. There is no need to re-implement it each time.

10
modules/core/doc/clustering.rst
View File

@@ -15,7 +15,15 @@ Finds centers of clusters and groups input samples around the clusters.
:param samples: Floating-point matrix of input samples, one row per sample.
:param data: Data for clustering.
:param data: Data for clustering. An array of N-Dimensional points with float coordinates is needed. Examples of this array can be:
* ``Mat points(count, 2, CV_32F);``
* ``Mat points(count, 1, CV_32FC2);``
* ``Mat points(1, count, CV_32FC2);``
* ``std::vector<cv::Point2f> points(sampleCount);``
:param cluster_count: Number of clusters to split the set by.

33
modules/core/doc/drawing_functions.rst
View File

@@ -361,6 +361,37 @@ The function ``line`` draws the line segment between ``pt1`` and ``pt2`` points
Antialiased lines are drawn using Gaussian filtering.
arrowedLine
----------------
Draws a arrow segment pointing from the first point to the second one.
.. ocv:function:: void arrowedLine(InputOutputArray img, Point pt1, Point pt2, const Scalar& color, int thickness=1, int lineType=8, int shift=0, double tipLength=0.1)
:param img: Image.
:param pt1: The point the arrow starts from.
:param pt2: The point the arrow points to.
:param color: Line color.
:param thickness: Line thickness.
:param lineType: Type of the line:
* **8** (or omitted) - 8-connected line.
* **4** - 4-connected line.
* **CV_AA** - antialiased line.
:param shift: Number of fractional bits in the point coordinates.
:param tipLength: The length of the arrow tip in relation to the arrow length
The function ``arrowedLine`` draws an arrow between ``pt1`` and ``pt2`` points in the image. See also :ocv:func:`line`.
LineIterator
------------
.. ocv:class:: LineIterator
@@ -585,7 +616,7 @@ Draws a text string.
:param font: ``CvFont`` structure initialized using :ocv:cfunc:`InitFont`.
:param fontFace: Font type. One of ``FONT_HERSHEY_SIMPLEX``, ``FONT_HERSHEY_PLAIN``, ``FONT_HERSHEY_DUPLEX``, ``FONT_HERSHEY_COMPLEX``, ``FONT_HERSHEY_TRIPLEX``, ``FONT_HERSHEY_COMPLEX_SMALL``, ``FONT_HERSHEY_SCRIPT_SIMPLEX``, or ``FONT_HERSHEY_SCRIPT_COMPLEX``,
where each of the font ID's can be combined with ``FONT_HERSHEY_ITALIC`` to get the slanted letters.
where each of the font ID's can be combined with ``FONT_ITALIC`` to get the slanted letters.
:param fontScale: Font scale factor that is multiplied by the font-specific base size.

2
modules/core/doc/dynamic_structures.rst
View File

@@ -164,7 +164,7 @@ field is negative (the most-significant bit, or MSB, of the field is set), and t
Initially the set and the free node list are empty. When a new node is requested from the set, it is taken from the list of free nodes, which is then updated. If the list appears to be empty, a new sequence block is allocated and all the nodes within the block are joined in the list of free nodes. Thus, the ``total``
field of the set is the total number of nodes both occupied and free. When an occupied node is released, it is added to the list of free nodes. The node released last will be occupied first.
``CvSet`` is used to represent graphs (:ocv:struct:`CvGraph`), sparse multi-dimensional arrays (:ocv:struct:`CvSparseMat`), and planar subdivisions (:ocv:struct:`CvSubdiv2D`).
``CvSet`` is used to represent graphs (:ocv:struct:`CvGraph`), sparse multi-dimensional arrays (:ocv:struct:`CvSparseMat`), and planar subdivisions (``CvSubdiv2D``).
CvSetElem
---------

3
modules/core/doc/intro.rst
View File

@@ -14,7 +14,8 @@ OpenCV has a modular structure, which means that the package includes several sh
* **calib3d** - basic multiple-view geometry algorithms, single and stereo camera calibration, object pose estimation, stereo correspondence algorithms, and elements of 3D reconstruction.
* **features2d** - salient feature detectors, descriptors, and descriptor matchers.
* **objdetect** - detection of objects and instances of the predefined classes (for example, faces, eyes, mugs, people, cars, and so on).
* **highgui** - an easy-to-use interface to video capturing, image and video codecs, as well as simple UI capabilities.
* **highgui** - an easy-to-use interface to simple UI capabilities.
* **videoio** - an easy-to-use interface to video capturing and video codecs.
* **gpu** - GPU-accelerated algorithms from different OpenCV modules.
* ... some other helper modules, such as FLANN and Google test wrappers, Python bindings, and others.

6
modules/core/doc/operations_on_arrays.rst
View File

@@ -1216,9 +1216,9 @@ gemm
----
Performs generalized matrix multiplication.
.. ocv:function:: void gemm( InputArray src1, InputArray src2, double alpha, InputArray src3, double gamma, OutputArray dst, int flags=0 )
.. ocv:function:: void gemm( InputArray src1, InputArray src2, double alpha, InputArray src3, double beta, OutputArray dst, int flags=0 )
.. ocv:pyfunction:: cv2.gemm(src1, src2, alpha, src3, gamma[, dst[, flags]]) -> dst
.. ocv:pyfunction:: cv2.gemm(src1, src2, alpha, src3, beta[, dst[, flags]]) -> dst
.. ocv:cfunction:: void cvGEMM( const CvArr* src1, const CvArr* src2, double alpha, const CvArr* src3, double beta, CvArr* dst, int tABC=0)
@@ -2065,7 +2065,7 @@ normalize
---------
Normalizes the norm or value range of an array.
.. ocv:function:: void normalize( InputArray src, OutputArray dst, double alpha=1, double beta=0, int norm_type=NORM_L2, int dtype=-1, InputArray mask=noArray() )
.. ocv:function:: void normalize( InputArray src, InputOutputArray dst, double alpha=1, double beta=0, int norm_type=NORM_L2, int dtype=-1, InputArray mask=noArray() )
.. ocv:function:: void normalize(const SparseMat& src, SparseMat& dst, double alpha, int normType)

20
modules/core/include/opencv2/core.hpp
View File

@@ -240,7 +240,7 @@ CV_EXPORTS_W void batchDistance(InputArray src1, InputArray src2,
bool crosscheck = false);
//! scales and shifts array elements so that either the specified norm (alpha) or the minimum (alpha) and maximum (beta) array values get the specified values
CV_EXPORTS_W void normalize( InputArray src, OutputArray dst, double alpha = 1, double beta = 0,
CV_EXPORTS_W void normalize( InputArray src, InputOutputArray dst, double alpha = 1, double beta = 0,
int norm_type = NORM_L2, int dtype = -1, InputArray mask = noArray());
//! scales and shifts array elements so that either the specified norm (alpha) or the minimum (alpha) and maximum (beta) array values get the specified values
@@ -389,7 +389,7 @@ CV_EXPORTS_W void patchNaNs(InputOutputArray a, double val = 0);
//! implements generalized matrix product algorithm GEMM from BLAS
CV_EXPORTS_W void gemm(InputArray src1, InputArray src2, double alpha,
InputArray src3, double gamma, OutputArray dst, int flags = 0);
InputArray src3, double beta, OutputArray dst, int flags = 0);
//! multiplies matrix by its transposition from the left or from the right
CV_EXPORTS_W void mulTransposed( InputArray src, OutputArray dst, bool aTa,
@@ -510,6 +510,10 @@ CV_EXPORTS_W void randShuffle(InputOutputArray dst, double iterFactor = 1., RNG*
CV_EXPORTS_W void line(InputOutputArray img, Point pt1, Point pt2, const Scalar& color,
int thickness = 1, int lineType = LINE_8, int shift = 0);
//! draws an arrow from pt1 to pt2 in the image
CV_EXPORTS_W void arrowedLine(InputOutputArray img, Point pt1, Point pt2, const Scalar& color,
int thickness=1, int line_type=8, int shift=0, double tipLength=0.1);
//! draws the rectangle outline or a solid rectangle with the opposite corners pt1 and pt2 in the image
CV_EXPORTS_W void rectangle(InputOutputArray img, Point pt1, Point pt2,
const Scalar& color, int thickness = 1,
@@ -953,12 +957,12 @@ public:
class CV_EXPORTS Formatter
{
public:
enum { FMT_MATLAB = 0,
FMT_CSV = 1,
FMT_PYTHON = 2,
FMT_NUMPY = 3,
FMT_C = 4,
FMT_DEFAULT = FMT_MATLAB
enum { FMT_DEFAULT = 0,
FMT_MATLAB = 1,
FMT_CSV = 2,
FMT_PYTHON = 3,
FMT_NUMPY = 4,
FMT_C = 5
};
virtual ~Formatter();

10
modules/core/include/opencv2/core/core_c.h
View File

@@ -1764,16 +1764,10 @@ CVAPI(int) cvGuiBoxReport( int status, const char* func_name, const char* err_ms
#define OPENCV_ERROR(status,func,context) \
cvError((status),(func),(context),__FILE__,__LINE__)
#define OPENCV_ERRCHK(func,context) \
{if (cvGetErrStatus() >= 0) \
{OPENCV_ERROR(CV_StsBackTrace,(func),(context));}}
#define OPENCV_ASSERT(expr,func,context) \
{if (! (expr)) \
{OPENCV_ERROR(CV_StsInternal,(func),(context));}}
#define OPENCV_RSTERR() (cvSetErrStatus(CV_StsOk))
#define OPENCV_CALL( Func ) \
{ \
Func; \
@@ -1800,10 +1794,6 @@ static char cvFuncName[] = Name
__CV_EXIT__; \
}
/* Simplified form of CV_ERROR */
#define CV_ERROR_FROM_CODE( code ) \
CV_ERROR( code, "" )
/*
CV_CHECK macro checks error status after CV (or IPL)
function call. If error detected, control will be transferred to the exit

1
modules/core/include/opencv2/core/cvdef.h
View File

@@ -244,6 +244,7 @@ typedef signed char schar;
/* fundamental constants */
#define CV_PI 3.1415926535897932384626433832795
#define CV_2PI 6.283185307179586476925286766559
#define CV_LOG2 0.69314718055994530941723212145818
/****************************************************************************************\

5
modules/core/include/opencv2/core/mat.hpp
View File

@@ -131,6 +131,7 @@ public:
virtual bool isSubmatrix(int i=-1) const;
virtual bool empty() const;
virtual void copyTo(const _OutputArray& arr) const;
virtual void copyTo(const _OutputArray& arr, const _InputArray & mask) const;
virtual size_t offset(int i=-1) const;
virtual size_t step(int i=-1) const;
bool isMat() const;
@@ -359,7 +360,7 @@ struct CV_EXPORTS UMatData
{
enum { COPY_ON_MAP=1, HOST_COPY_OBSOLETE=2,
DEVICE_COPY_OBSOLETE=4, TEMP_UMAT=8, TEMP_COPIED_UMAT=24,
USER_ALLOCATED=32 };
USER_ALLOCATED=32, DEVICE_MEM_MAPPED=64};
UMatData(const MatAllocator* allocator);
~UMatData();
@@ -369,11 +370,13 @@ struct CV_EXPORTS UMatData
bool hostCopyObsolete() const;
bool deviceCopyObsolete() const;
bool deviceMemMapped() const;
bool copyOnMap() const;
bool tempUMat() const;
bool tempCopiedUMat() const;
void markHostCopyObsolete(bool flag);
void markDeviceCopyObsolete(bool flag);
void markDeviceMemMapped(bool flag);
const MatAllocator* prevAllocator;
const MatAllocator* currAllocator;

9
modules/core/include/opencv2/core/mat.inl.hpp
View File

@@ -3350,10 +3350,19 @@ size_t UMat::total() const
inline bool UMatData::hostCopyObsolete() const { return (flags & HOST_COPY_OBSOLETE) != 0; }
inline bool UMatData::deviceCopyObsolete() const { return (flags & DEVICE_COPY_OBSOLETE) != 0; }
inline bool UMatData::deviceMemMapped() const { return (flags & DEVICE_MEM_MAPPED) != 0; }
inline bool UMatData::copyOnMap() const { return (flags & COPY_ON_MAP) != 0; }
inline bool UMatData::tempUMat() const { return (flags & TEMP_UMAT) != 0; }
inline bool UMatData::tempCopiedUMat() const { return (flags & TEMP_COPIED_UMAT) == TEMP_COPIED_UMAT; }
inline void UMatData::markDeviceMemMapped(bool flag)
{
if(flag)
flags |= DEVICE_MEM_MAPPED;
else
flags &= ~DEVICE_MEM_MAPPED;
}
inline void UMatData::markHostCopyObsolete(bool flag)
{
if(flag)

12
modules/core/include/opencv2/core/ocl.hpp
View File

@@ -46,12 +46,12 @@
namespace cv { namespace ocl {
CV_EXPORTS bool haveOpenCL();
CV_EXPORTS bool useOpenCL();
CV_EXPORTS bool haveAmdBlas();
CV_EXPORTS bool haveAmdFft();
CV_EXPORTS void setUseOpenCL(bool flag);
CV_EXPORTS void finish();
CV_EXPORTS_W bool haveOpenCL();
CV_EXPORTS_W bool useOpenCL();
CV_EXPORTS_W bool haveAmdBlas();
CV_EXPORTS_W bool haveAmdFft();
CV_EXPORTS_W void setUseOpenCL(bool flag);
CV_EXPORTS_W void finish();
class CV_EXPORTS Context;
class CV_EXPORTS Device;

21
modules/core/include/opencv2/core/private.hpp
View File

@@ -210,14 +210,12 @@ CV_EXPORTS void scalarToRawData(const cv::Scalar& s, void* buf, int type, int un
\****************************************************************************************/
#ifdef HAVE_IPP
# ifdef HAVE_IPP_ICV_ONLY
# include "ipp_redefine.h"
# include "ippicv.h"
# else
# include "ipp.h"
# endif
# include "ipp.h"
# define IPP_VERSION_X100 (IPP_VERSION_MAJOR * 100 + IPP_VERSION_MINOR)
#define IPP_ALIGN 32 // required for AVX optimization
#define setIppErrorStatus() cv::ipp::setIppStatus(-1, CV_Func, __FILE__, __LINE__)
static inline IppiSize ippiSize(int width, int height)
@@ -241,6 +239,17 @@ static inline IppiBorderType ippiGetBorderType(int borderTypeNI)
borderTypeNI == cv::BORDER_REFLECT ? ippBorderMirrorR : (IppiBorderType)-1;
}
static inline IppDataType ippiGetDataType(int depth)
{
return depth == CV_8U ? ipp8u :
depth == CV_8S ? ipp8s :
depth == CV_16U ? ipp16u :
depth == CV_16S ? ipp16s :
depth == CV_32S ? ipp32s :
depth == CV_32F ? ipp32f :
depth == CV_64F ? ipp64f : (IppDataType)-1;
}
#else
# define IPP_VERSION_X100 0
#endif

111
modules/core/perf/opencl/perf_arithm.cpp
View File

@@ -308,6 +308,23 @@ OCL_PERF_TEST_P(TransposeFixture, Transpose, ::testing::Combine(
SANITY_CHECK(dst);
}
OCL_PERF_TEST_P(TransposeFixture, TransposeInplace, ::testing::Combine(
OCL_PERF_ENUM(Size(640, 640), Size(1280, 1280), Size(2160, 2160)), OCL_TEST_TYPES_134))
{
const Size_MatType_t params = GetParam();
const Size srcSize = get<0>(params);
const int type = get<1>(params);
checkDeviceMaxMemoryAllocSize(srcSize, type);
UMat src(srcSize, type);
declare.in(src, WARMUP_RNG).out(src, WARMUP_NONE);
OCL_TEST_CYCLE() cv::transpose(src, src);
SANITY_CHECK_NOTHING();
}
///////////// Flip ////////////////////////
enum
@@ -591,12 +608,12 @@ OCL_PERF_TEST_P(PowFixture, Pow, ::testing::Combine(
checkDeviceMaxMemoryAllocSize(srcSize, type);
UMat src(srcSize, type), dst(srcSize, type);
randu(src, -100, 100);
randu(src, 0, 100);
declare.in(src).out(dst);
OCL_TEST_CYCLE() cv::pow(src, -2.0, dst);
OCL_TEST_CYCLE() cv::pow(src, 2.17, dst);
SANITY_CHECK(dst, 1e-6, ERROR_RELATIVE);
SANITY_CHECK(dst, 1.5e-6, ERROR_RELATIVE);
}
///////////// AddWeighted////////////////////////
@@ -701,6 +718,36 @@ OCL_PERF_TEST_P(MeanStdDevFixture, MeanStdDev,
SANITY_CHECK(stddev3, eps, ERROR_RELATIVE);
}
OCL_PERF_TEST_P(MeanStdDevFixture, MeanStdDevWithMask,
::testing::Combine(OCL_PERF_ENUM(OCL_SIZE_1, OCL_SIZE_2, OCL_SIZE_3),
OCL_TEST_TYPES_134))
{
const Size_MatType_t params = GetParam();
const Size srcSize = get<0>(params);
const int type = get<1>(params);
const double eps = 2e-5;
checkDeviceMaxMemoryAllocSize(srcSize, type);
UMat src(srcSize, type), mask(srcSize, CV_8UC1);
Scalar mean, stddev;
declare.in(src, mask, WARMUP_RNG);
OCL_TEST_CYCLE() cv::meanStdDev(src, mean, stddev, mask);
double mean0 = mean[0], mean1 = mean[1], mean2 = mean[2], mean3 = mean[3];
double stddev0 = stddev[0], stddev1 = stddev[1], stddev2 = stddev[2], stddev3 = stddev[3];
SANITY_CHECK(mean0, eps, ERROR_RELATIVE);
SANITY_CHECK(mean1, eps, ERROR_RELATIVE);
SANITY_CHECK(mean2, eps, ERROR_RELATIVE);
SANITY_CHECK(mean3, eps, ERROR_RELATIVE);
SANITY_CHECK(stddev0, eps, ERROR_RELATIVE);
SANITY_CHECK(stddev1, eps, ERROR_RELATIVE);
SANITY_CHECK(stddev2, eps, ERROR_RELATIVE);
SANITY_CHECK(stddev3, eps, ERROR_RELATIVE);
}
///////////// Norm ////////////////////////
CV_ENUM(NormType, NORM_INF, NORM_L1, NORM_L2)
@@ -708,6 +755,26 @@ CV_ENUM(NormType, NORM_INF, NORM_L1, NORM_L2)
typedef std::tr1::tuple<Size, MatType, NormType> NormParams;
typedef TestBaseWithParam<NormParams> NormFixture;
OCL_PERF_TEST_P(NormFixture, Norm1Arg,
::testing::Combine(OCL_PERF_ENUM(OCL_SIZE_1, OCL_SIZE_2, OCL_SIZE_3),
OCL_TEST_TYPES_134, NormType::all()))
{
const NormParams params = GetParam();
const Size srcSize = get<0>(params);
const int type = get<1>(params);
const int normType = get<2>(params);
checkDeviceMaxMemoryAllocSize(srcSize, type);
UMat src1(srcSize, type);
double res;
declare.in(src1, WARMUP_RNG);
OCL_TEST_CYCLE() res = cv::norm(src1, normType);
SANITY_CHECK(res, 1e-5, ERROR_RELATIVE);
}
OCL_PERF_TEST_P(NormFixture, Norm,
::testing::Combine(OCL_PERF_ENUM(OCL_SIZE_1, OCL_SIZE_2, OCL_SIZE_3),
OCL_TEST_TYPES_134, NormType::all()))
@@ -728,6 +795,26 @@ OCL_PERF_TEST_P(NormFixture, Norm,
SANITY_CHECK(res, 1e-5, ERROR_RELATIVE);
}
OCL_PERF_TEST_P(NormFixture, NormRel,
::testing::Combine(OCL_PERF_ENUM(OCL_SIZE_1, OCL_SIZE_2, OCL_SIZE_3),
OCL_TEST_TYPES_134, NormType::all()))
{
const NormParams params = GetParam();
const Size srcSize = get<0>(params);
const int type = get<1>(params);
const int normType = get<2>(params);
checkDeviceMaxMemoryAllocSize(srcSize, type);
UMat src1(srcSize, type), src2(srcSize, type);
double res;
declare.in(src1, src2, WARMUP_RNG);
OCL_TEST_CYCLE() res = cv::norm(src1, src2, normType | cv::NORM_RELATIVE);
SANITY_CHECK(res, 1e-5, ERROR_RELATIVE);
}
///////////// UMat::dot ////////////////////////
typedef Size_MatType UMatDotFixture;
@@ -860,6 +947,24 @@ OCL_PERF_TEST_P(NormalizeFixture, Normalize,
SANITY_CHECK(dst, 5e-2);
}
OCL_PERF_TEST_P(NormalizeFixture, NormalizeWithMask,
::testing::Combine(OCL_TEST_SIZES, OCL_PERF_ENUM(CV_8UC1, CV_32FC1),
NormalizeModes::all()))
{
const NormalizeParams params = GetParam();
const Size srcSize = get<0>(params);
const int type = get<1>(params), mode = get<2>(params);
checkDeviceMaxMemoryAllocSize(srcSize, type);
UMat src(srcSize, type), mask(srcSize, CV_8UC1), dst(srcSize, type);
declare.in(src, mask, WARMUP_RNG).out(dst);
OCL_TEST_CYCLE() cv::normalize(src, dst, 10, 110, mode, -1, mask);
SANITY_CHECK(dst, 5e-2);
}
///////////// ConvertScaleAbs ////////////////////////
typedef Size_MatType ConvertScaleAbsFixture;

37
modules/core/perf/opencl/perf_dxt.cpp
View File

@@ -54,23 +54,42 @@ namespace ocl {
///////////// dft ////////////////////////
typedef tuple<Size, int> DftParams;
enum OCL_FFT_TYPE
{
R2R = 0,
C2R = 1,
R2C = 2,
C2C = 3
};
typedef tuple<OCL_FFT_TYPE, Size, int> DftParams;
typedef TestBaseWithParam<DftParams> DftFixture;
OCL_PERF_TEST_P(DftFixture, Dft, ::testing::Combine(Values(OCL_SIZE_1, OCL_SIZE_2, OCL_SIZE_3),
Values((int)DFT_ROWS, (int)DFT_SCALE, (int)DFT_INVERSE,
(int)DFT_INVERSE | DFT_SCALE, (int)DFT_ROWS | DFT_INVERSE)))
OCL_PERF_TEST_P(DftFixture, Dft, ::testing::Combine(Values(C2C, R2R, C2R, R2C),
Values(OCL_SIZE_1, OCL_SIZE_2, OCL_SIZE_3, Size(512, 512), Size(1024, 1024), Size(2048, 2048)),
Values((int) 0, (int)DFT_ROWS, (int)DFT_SCALE, (int)DFT_INVERSE,
(int)DFT_INVERSE | DFT_SCALE, (int)DFT_ROWS | DFT_INVERSE)))
{
const DftParams params = GetParam();
const Size srcSize = get<0>(params);
const int flags = get<1>(params);
const int dft_type = get<0>(params);
const Size srcSize = get<1>(params);
int flags = get<2>(params);
UMat src(srcSize, CV_32FC2), dst(srcSize, CV_32FC2);
int in_cn, out_cn;
switch (dft_type)
{
case R2R: flags |= cv::DFT_REAL_OUTPUT; in_cn = 1; out_cn = 1; break;
case C2R: flags |= cv::DFT_REAL_OUTPUT; in_cn = 2; out_cn = 2; break;
case R2C: flags |= cv::DFT_COMPLEX_OUTPUT; in_cn = 1; out_cn = 2; break;
case C2C: flags |= cv::DFT_COMPLEX_OUTPUT; in_cn = 2; out_cn = 2; break;
}
UMat src(srcSize, CV_MAKE_TYPE(CV_32F, in_cn)), dst(srcSize, CV_MAKE_TYPE(CV_32F, out_cn));
declare.in(src, WARMUP_RNG).out(dst);
OCL_TEST_CYCLE() cv::dft(src, dst, flags | DFT_COMPLEX_OUTPUT);
OCL_TEST_CYCLE() cv::dft(src, dst, flags);
SANITY_CHECK(dst, 1e-3);
SANITY_CHECK(dst, 1e-5, ERROR_RELATIVE);
}
///////////// MulSpectrums ////////////////////////

24
modules/core/perf/opencl/perf_matop.cpp
View File

@@ -122,6 +122,30 @@ OCL_PERF_TEST_P(CopyToFixture, CopyToWithMask,
SANITY_CHECK(dst);
}
OCL_PERF_TEST_P(CopyToFixture, CopyToWithMaskUninit,
::testing::Combine(OCL_PERF_ENUM(OCL_SIZE_1, OCL_SIZE_2, OCL_SIZE_3), OCL_TEST_TYPES))
{
const Size_MatType_t params = GetParam();
const Size srcSize = get<0>(params);
const int type = get<1>(params);
checkDeviceMaxMemoryAllocSize(srcSize, type);
UMat src(srcSize, type), dst, mask(srcSize, CV_8UC1);
declare.in(src, mask, WARMUP_RNG);
for ( ; next(); )
{
dst.release();
startTimer();
src.copyTo(dst, mask);
cv::ocl::finish();
stopTimer();
}
SANITY_CHECK(dst);
}
} } // namespace cvtest::ocl
#endif // HAVE_OPENCL

Some files were not shown because too many files have changed in this diff Show More