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

use perf test replace performance sample

Browse Source
This commit is contained in:
yao
2013-03-27 12:04:48 +08:00
parent 55c9a7c87d
commit 5539e85a11
26 changed files with 3791 additions and 14305 deletions
Download Patch File Download Diff File Expand all files Collapse all files

150
modules/ocl/perf/perf_hog.cpp
View File

@@ -15,7 +15,7 @@
// Third party copyrights are property of their respective owners.
//
// @Authors
// Peng Xiao, pengxiao@multicorewareinc.com
// Fangfang Bai, fangfang@multicorewareinc.com
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
@@ -42,125 +42,47 @@
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "precomp.hpp"
#include <iomanip>
#ifdef HAVE_OPENCL
using namespace cv;
using namespace cv::ocl;
using namespace cvtest;
using namespace testing;
using namespace std;
extern std::string workdir;
#ifndef MWC_TEST_UTILITY
#define MWC_TEST_UTILITY
// Param class
#ifndef IMPLEMENT_PARAM_CLASS
#define IMPLEMENT_PARAM_CLASS(name, type) \
class name \
{ \
public: \
name ( type arg = type ()) : val_(arg) {} \
operator type () const {return val_;} \
private: \
type val_; \
}; \
inline void PrintTo( name param, std::ostream* os) \
{ \
*os << #name << "(" << testing::PrintToString(static_cast< type >(param)) << ")"; \
}
#endif // IMPLEMENT_PARAM_CLASS
#endif // MWC_TEST_UTILITY
IMPLEMENT_PARAM_CLASS(WinSizw48, bool);
PARAM_TEST_CASE(HOG, WinSizw48, bool)
///////////// HOG////////////////////////
TEST(HOG)
{
bool is48;
vector<float> detector;
virtual void SetUp()
Mat src = imread(abspath("road.png"), cv::IMREAD_GRAYSCALE);
if (src.empty())
{
is48 = GET_PARAM(0);
if(is48)
{
detector = cv::ocl::HOGDescriptor::getPeopleDetector48x96();
}
else
{
detector = cv::ocl::HOGDescriptor::getPeopleDetector64x128();
}
}
};
TEST_P(HOG, Performance)
{
cv::Mat img = readImage(workdir + "lena.jpg", cv::IMREAD_GRAYSCALE);
ASSERT_FALSE(img.empty());
// define HOG related arguments
float scale = 1.05f;
//int nlevels = 13;
int gr_threshold = 8;
float hit_threshold = 1.4f;
//bool hit_threshold_auto = true;
int win_width = is48 ? 48 : 64;
int win_stride_width = 8;
int win_stride_height = 8;
bool gamma_corr = true;
Size win_size(win_width, win_width * 2); //(64, 128) or (48, 96)
Size win_stride(win_stride_width, win_stride_height);
cv::ocl::HOGDescriptor gpu_hog(win_size, Size(16, 16), Size(8, 8), Size(8, 8), 9,
cv::ocl::HOGDescriptor::DEFAULT_WIN_SIGMA, 0.2, gamma_corr,
cv::ocl::HOGDescriptor::DEFAULT_NLEVELS);
gpu_hog.setSVMDetector(detector);
double totalgputick = 0;
double totalgputick_kernel = 0;
double t1 = 0;
double t2 = 0;
for(int j = 0; j < LOOP_TIMES + 1; j ++)
{
t1 = (double)cvGetTickCount();//gpu start1
ocl::oclMat d_src(img);//upload
t2 = (double)cvGetTickCount(); //kernel
vector<Rect> found;
gpu_hog.detectMultiScale(d_src, found, hit_threshold, win_stride,
Size(0, 0), scale, gr_threshold);
t2 = (double)cvGetTickCount() - t2;//kernel
// no download time for HOG
t1 = (double)cvGetTickCount() - t1;//gpu end1
if(j == 0)
continue;
totalgputick = t1 + totalgputick;
totalgputick_kernel = t2 + totalgputick_kernel;
throw runtime_error("can't open road.png");
}
cout << "average gpu runtime is " << totalgputick / ((double)cvGetTickFrequency()* LOOP_TIMES * 1000.) << "ms" << endl;
cout << "average gpu runtime without data transfer is " << totalgputick_kernel / ((double)cvGetTickFrequency()* LOOP_TIMES * 1000.) << "ms" << endl;
}
cv::HOGDescriptor hog;
hog.setSVMDetector(hog.getDefaultPeopleDetector());
std::vector<cv::Rect> found_locations;
INSTANTIATE_TEST_CASE_P(GPU_ObjDetect, HOG, testing::Combine(testing::Values(WinSizw48(false), WinSizw48(true)), testing::Values(false)));
SUBTEST << 768 << 'x' << 576 << "; road.png";
#endif //Have opencl
hog.detectMultiScale(src, found_locations);
CPU_ON;
hog.detectMultiScale(src, found_locations);
CPU_OFF;
cv::ocl::HOGDescriptor ocl_hog;
ocl_hog.setSVMDetector(ocl_hog.getDefaultPeopleDetector());
ocl::oclMat d_src;
d_src.upload(src);
WARMUP_ON;
ocl_hog.detectMultiScale(d_src, found_locations);
WARMUP_OFF;
GPU_ON;
ocl_hog.detectMultiScale(d_src, found_locations);
;
GPU_OFF;
GPU_FULL_ON;
d_src.upload(src);
ocl_hog.detectMultiScale(d_src, found_locations);
GPU_FULL_OFF;
}