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

update some of the functions in ocl module to the latest version

Browse Source
This commit is contained in:
yao
2012-09-03 17:03:37 +08:00
parent 3fb3851c7a
commit 0fdb55a54d
24 changed files with 2931 additions and 1728 deletions
Download Patch File Download Diff File Expand all files Collapse all files

117
modules/ocl/src/hog.cpp
View File

@@ -119,6 +119,8 @@ namespace cv { namespace ocl { namespace device
float angle_scale, cv::ocl::oclMat& grad, cv::ocl::oclMat& qangle, bool correct_gamma);
void compute_gradients_8UC4(int height, int width, const cv::ocl::oclMat& img,
float angle_scale, cv::ocl::oclMat& grad, cv::ocl::oclMat& qangle, bool correct_gamma);
void resize( const oclMat &src, oclMat &dst, const Size sz);
}
}}}
@@ -150,6 +152,8 @@ cv::ocl::HOGDescriptor::HOGDescriptor(Size win_size_, Size block_size_, Size blo
cv::Size blocks_per_win = numPartsWithin(win_size, block_size, block_stride);
hog::set_up_constants(nbins, block_stride.width, block_stride.height, blocks_per_win.width, blocks_per_win.height);
effect_size = Size(0, 0);
}
size_t cv::ocl::HOGDescriptor::getDescriptorSize() const
@@ -199,22 +203,37 @@ void cv::ocl::HOGDescriptor::setSVMDetector(const vector<float>& _detector)
CV_Assert(checkDetectorSize());
}
void cv::ocl::HOGDescriptor::init_buffer(const oclMat& img, Size win_stride)
{
if (!image_scale.empty())
return;
if (effect_size == Size(0, 0))
effect_size = img.size();
grad.create(img.size(), CV_32FC2);
qangle.create(img.size(), CV_8UC2);
const size_t block_hist_size = getBlockHistogramSize();
const Size blocks_per_img = numPartsWithin(img.size(), block_size, block_stride);
block_hists.create(1, static_cast<int>(block_hist_size * blocks_per_img.area()), CV_32F);
Size wins_per_img = numPartsWithin(img.size(), win_size, win_stride);
labels.create(1, wins_per_img.area(), CV_8U);
}
void cv::ocl::HOGDescriptor::computeGradient(const oclMat& img, oclMat& grad, oclMat& qangle)
{
CV_Assert(img.type() == CV_8UC1 || img.type() == CV_8UC4);
grad.create(img.size(), CV_32FC2);
qangle.create(img.size(), CV_8UC2);
float angleScale = (float)(nbins / CV_PI);
switch (img.type())
{
case CV_8UC1:
hog::compute_gradients_8UC1(img.rows, img.cols, img, angleScale, grad, qangle, gamma_correction);
hog::compute_gradients_8UC1(effect_size.height, effect_size.width, img, angleScale, grad, qangle, gamma_correction);
break;
case CV_8UC4:
hog::compute_gradients_8UC4(img.rows, img.cols, img, angleScale, grad, qangle, gamma_correction);
hog::compute_gradients_8UC4(effect_size.height, effect_size.width, img, angleScale, grad, qangle, gamma_correction);
break;
}
}
@@ -224,14 +243,11 @@ void cv::ocl::HOGDescriptor::computeBlockHistograms(const oclMat& img)
{
computeGradient(img, grad, qangle);
size_t block_hist_size = getBlockHistogramSize();
Size blocks_per_img = numPartsWithin(img.size(), block_size, block_stride);
hog::compute_hists(nbins, block_stride.width, block_stride.height, effect_size.height, effect_size.width,
grad, qangle, (float)getWinSigma(), block_hists);
block_hists.create(1, static_cast<int>(block_hist_size * blocks_per_img.area()), CV_32F);
hog::compute_hists(nbins, block_stride.width, block_stride.height, img.rows, img.cols, grad, qangle, (float)getWinSigma(), block_hists);
hog::normalize_hists(nbins, block_stride.width, block_stride.height, img.rows, img.cols, block_hists, (float)threshold_L2hys);
hog::normalize_hists(nbins, block_stride.width, block_stride.height, effect_size.height, effect_size.width,
block_hists, (float)threshold_L2hys);
}
@@ -239,11 +255,13 @@ void cv::ocl::HOGDescriptor::getDescriptors(const oclMat& img, Size win_stride,
{
CV_Assert(win_stride.width % block_stride.width == 0 && win_stride.height % block_stride.height == 0);
init_buffer(img, win_stride);
computeBlockHistograms(img);
const size_t block_hist_size = getBlockHistogramSize();
Size blocks_per_win = numPartsWithin(win_size, block_size, block_stride);
Size wins_per_img = numPartsWithin(img.size(), win_size, win_stride);
Size wins_per_img = numPartsWithin(effect_size, win_size, win_stride);
descriptors.create(wins_per_img.area(), static_cast<int>(blocks_per_win.area() * block_hist_size), CV_32F);
@@ -251,11 +269,11 @@ void cv::ocl::HOGDescriptor::getDescriptors(const oclMat& img, Size win_stride,
{
case DESCR_FORMAT_ROW_BY_ROW:
hog::extract_descrs_by_rows(win_size.height, win_size.width, block_stride.height, block_stride.width,
win_stride.height, win_stride.width, img.rows, img.cols, block_hists, descriptors);
win_stride.height, win_stride.width, effect_size.height, effect_size.width, block_hists, descriptors);
break;
case DESCR_FORMAT_COL_BY_COL:
hog::extract_descrs_by_cols(win_size.height, win_size.width, block_stride.height, block_stride.width,
win_stride.height, win_stride.width, img.rows, img.cols, block_hists, descriptors);
win_stride.height, win_stride.width, effect_size.height, effect_size.width, block_hists, descriptors);
break;
default:
CV_Error(CV_StsBadArg, "Unknown descriptor format");
@@ -272,22 +290,21 @@ void cv::ocl::HOGDescriptor::detect(const oclMat& img, vector<Point>& hits, doub
if (detector.empty())
return;
computeBlockHistograms(img);
if (win_stride == Size())
win_stride = block_stride;
else
CV_Assert(win_stride.width % block_stride.width == 0 && win_stride.height % block_stride.height == 0);
init_buffer(img, win_stride);
Size wins_per_img = numPartsWithin(img.size(), win_size, win_stride);
labels.create(1, wins_per_img.area(), CV_8U);
computeBlockHistograms(img);
hog::classify_hists(win_size.height, win_size.width, block_stride.height, block_stride.width,
win_stride.height, win_stride.width, img.rows, img.cols, block_hists,
win_stride.height, win_stride.width, effect_size.height, effect_size.width, block_hists,
detector, (float)free_coef, (float)hit_threshold, labels);
labels.download(labels_host);
unsigned char* vec = labels_host.ptr();
Size wins_per_img = numPartsWithin(effect_size, win_size, win_stride);
for (int i = 0; i < wins_per_img.area(); i++)
{
int y = i / wins_per_img.width;
@@ -303,6 +320,7 @@ void cv::ocl::HOGDescriptor::detectMultiScale(const oclMat& img, vector<Rect>& f
Size win_stride, Size padding, double scale0, int group_threshold)
{
CV_Assert(img.type() == CV_8UC1 || img.type() == CV_8UC4);
CV_Assert(scale0 > 1);
vector<double> level_scale;
double scale = 1.;
@@ -318,27 +336,30 @@ void cv::ocl::HOGDescriptor::detectMultiScale(const oclMat& img, vector<Rect>& f
}
levels = std::max(levels, 1);
level_scale.resize(levels);
image_scales.resize(levels);
std::vector<Rect> all_candidates;
vector<Point> locations;
if (win_stride == Size())
win_stride = block_stride;
else
CV_Assert(win_stride.width % block_stride.width == 0 && win_stride.height % block_stride.height == 0);
init_buffer(img, win_stride);
image_scale.create(img.size(), img.type());
for (size_t i = 0; i < level_scale.size(); i++)
{
scale = level_scale[i];
Size sz(cvRound(img.cols / scale), cvRound(img.rows / scale));
oclMat smaller_img;
if (sz == img.size())
smaller_img = img;
effect_size = Size(cvRound(img.cols / scale), cvRound(img.rows / scale));
if (effect_size == img.size())
{
detect(img, locations, hit_threshold, win_stride, padding);
}
else
{
image_scales[i].create(sz, img.type());
resize(img, image_scales[i], image_scales[i].size(), 0, 0, INTER_LINEAR);
smaller_img = image_scales[i];
hog::resize( img, image_scale, effect_size);
detect(image_scale, locations, hit_threshold, win_stride, padding);
}
detect(smaller_img, locations, hit_threshold, win_stride, padding);
Size scaled_win_size(cvRound(win_size.width * scale), cvRound(win_size.height * scale));
for (size_t j = 0; j < locations.size(); j++)
all_candidates.push_back(Rect(Point2d((CvPoint)locations[j]) * scale, scaled_win_size));
@@ -1784,4 +1805,36 @@ void cv::ocl::device::hog::compute_gradients_8UC4(int height, int width, const c
openCLExecuteKernel(clCxt, &objdetect_hog, kernelName, globalThreads, localThreads, args, -1, -1);
}
void cv::ocl::device::hog::resize( const oclMat &src, oclMat &dst, const Size sz)
{
CV_Assert( (src.channels() == dst.channels()) );
Context *clCxt = Context::getContext();
string kernelName = (src.type() == CV_8UC1) ? "resize_8UC1_kernel" : "resize_8UC4_kernel";
size_t blkSizeX = 16, blkSizeY = 16;
size_t glbSizeX = sz.width % blkSizeX == 0 ? sz.width : (sz.width / blkSizeX + 1) * blkSizeX;
size_t glbSizeY = sz.height % blkSizeY == 0 ? sz.height : (sz.height / blkSizeY + 1) * blkSizeY;
size_t globalThreads[3] = {glbSizeX, glbSizeY, 1};
size_t localThreads[3] = {blkSizeX, blkSizeY, 1};
float ifx = (float)src.cols / sz.width;
float ify = (float)src.rows / sz.height;
vector< pair<size_t, const void *> > args;
args.push_back( make_pair(sizeof(cl_mem), (void *)&dst.data));
args.push_back( make_pair(sizeof(cl_mem), (void *)&src.data));
args.push_back( make_pair(sizeof(cl_int), (void *)&dst.offset));
args.push_back( make_pair(sizeof(cl_int), (void *)&src.offset));
args.push_back( make_pair(sizeof(cl_int), (void *)&dst.step));
args.push_back( make_pair(sizeof(cl_int), (void *)&src.step));
args.push_back( make_pair(sizeof(cl_int), (void *)&src.cols));
args.push_back( make_pair(sizeof(cl_int), (void *)&src.rows));
args.push_back( make_pair(sizeof(cl_int), (void *)&sz.width));
args.push_back( make_pair(sizeof(cl_int), (void *)&sz.height));
args.push_back( make_pair(sizeof(cl_float), (void *)&ifx));
args.push_back( make_pair(sizeof(cl_float), (void *)&ify));
openCLExecuteKernel(clCxt, &objdetect_hog, kernelName, globalThreads, localThreads, args, -1, -1);
}
#endif