Skip to content

O
opencv
Commit dab78c26 authored by Vadim Pisarevsky's avatar Vadim Pisarevsky
Browse files
Options

added reasonable test for MSER (including coverage for… 

added reasonable test for MSER (including coverage for http://code.opencv.org/issues/4273, http://code.opencv.org/issues/1723 and http://code.opencv.org/issues/756); also, added some "in-progress" info to the Features2d_Feature2d.no_crash test.
parent fede94e9
Hide whitespace changes
Inline Side-by-side
Showing with 100 additions and 143 deletions
modules/features2d/test/test_descriptors_regression.cpp
View file @ dab78c26
......@@ -390,45 +390,52 @@ TEST( Features2d_Feature2d, no_crash )
size_t i, n = fnames.size();
vector<KeyPoint> keypoints;
Mat descriptors;
orb->setMaxFeatures(5000);
for( i = 0; i < n; i++ )
{
printf("%d. image: %s:\n", (int)i, fnames[i].c_str());
if( strstr(fnames[i].c_str(), "MP.png") != 0 )
continue;
bool checkCount = strstr(fnames[i].c_str(), "templ.png") == 0;
Mat img = imread(fnames[i], -1);
printf("\tAKAZE ... "); fflush(stdout);
akaze->detectAndCompute(img, noArray(), keypoints, descriptors);
printf("(%d keypoints) ", (int)keypoints.size()); fflush(stdout);
if( checkCount )
{
ASSERT_GT((int)keypoints.size(), 0);
EXPECT_GT((int)keypoints.size(), 0);
}
ASSERT_EQ(descriptors.rows, (int)keypoints.size());
printf("ok\n");
printf("\tKAZE ... "); fflush(stdout);
kaze->detectAndCompute(img, noArray(), keypoints, descriptors);
printf("(%d keypoints) ", (int)keypoints.size()); fflush(stdout);
if( checkCount )
{
ASSERT_GT((int)keypoints.size(), 0);
EXPECT_GT((int)keypoints.size(), 0);
}
ASSERT_EQ(descriptors.rows, (int)keypoints.size());
printf("ok\n");
printf("\tORB ... "); fflush(stdout);
orb->detectAndCompute(img, noArray(), keypoints, descriptors);
printf("(%d keypoints) ", (int)keypoints.size()); fflush(stdout);
if( checkCount )
{
ASSERT_GT((int)keypoints.size(), 0);
EXPECT_GT((int)keypoints.size(), 0);
}
ASSERT_EQ(descriptors.rows, (int)keypoints.size());
printf("ok\n");
printf("\tBRISK ... "); fflush(stdout);
brisk->detectAndCompute(img, noArray(), keypoints, descriptors);
printf("(%d keypoints) ", (int)keypoints.size()); fflush(stdout);
if( checkCount )
{
ASSERT_GT((int)keypoints.size(), 0);
EXPECT_GT((int)keypoints.size(), 0);
}
ASSERT_EQ(descriptors.rows, (int)keypoints.size());
printf("ok\n");
......
modules/features2d/test/test_mser.cpp
View file @ dab78c26
......@@ -41,171 +41,121 @@
//M*/
#include "test_precomp.hpp"
#include "opencv2/imgproc/imgproc_c.h"
#if 0
#include "opencv2/highgui.hpp"
#include <vector>
#include <string>
using namespace std;
using namespace cv;
class CV_MserTest : public cvtest::BaseTest
{
public:
CV_MserTest();
protected:
void run(int);
int LoadBoxes(const char* path, vector<CvBox2D>& boxes);
int SaveBoxes(const char* path, const vector<CvBox2D>& boxes);
int CompareBoxes(const vector<CvBox2D>& boxes1,const vector<CvBox2D>& boxes2, float max_rel_diff = 0.01f);
};
CV_MserTest::CV_MserTest()
{
}
#undef RENDER_MSERS
#define RENDER_MSERS 0
int CV_MserTest::LoadBoxes(const char* path, vector<CvBox2D>& boxes)
#if defined RENDER_MSERS && RENDER_MSERS
static void renderMSERs(const Mat& gray, Mat& img, const vector<vector<Point> >& msers)
{
boxes.clear();
FILE* f = fopen(path,"r");
if (f==NULL)
cvtColor(gray, img, COLOR_GRAY2BGR);
RNG rng((uint64)1749583);
for( int i = 0; i < (int)msers.size(); i++ )
{
return 0;
uchar b = rng.uniform(0, 256);
uchar g = rng.uniform(0, 256);
uchar r = rng.uniform(0, 256);
Vec3b color(b, g, r);
const Point* pt = &msers[i][0];
size_t j, n = msers[i].size();
for( j = 0; j < n; j++ )
img.at<Vec3b>(pt[j]) = color;
}
while (!feof(f))
{
CvBox2D box;
int values_read = fscanf(f,"%f,%f,%f,%f,%f\n",&box.angle,&box.center.x,&box.center.y,&box.size.width,&box.size.height);
CV_Assert(values_read == 5);
boxes.push_back(box);
}
fclose(f);
return 1;
}
#endif
int CV_MserTest::SaveBoxes(const char* path, const vector<CvBox2D>& boxes)
TEST(Features2d_MSER, cases)
{
FILE* f = fopen(path,"w");
if (f==NULL)
{
return 0;
}
for (int i=0;i<(int)boxes.size();i++)
uchar buf[] =
{
fprintf(f,"%f,%f,%f,%f,%f\n",boxes[i].angle,boxes[i].center.x,boxes[i].center.y,boxes[i].size.width,boxes[i].size.height);
}
fclose(f);
return 1;
}
255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
255, 255, 255, 255, 255, 0, 0, 0, 0, 255, 255, 255, 255, 255, 255, 255, 255, 255, 0, 0, 0, 0, 255, 255, 255, 255,
255, 255, 255, 255, 255, 0, 0, 0, 0, 0, 255, 255, 255, 255, 255, 255, 255, 255, 0, 0, 0, 0, 255, 255, 255, 255,
255, 255, 255, 255, 255, 0, 0, 0, 0, 0, 255, 255, 255, 255, 255, 255, 255, 255, 0, 0, 0, 0, 255, 255, 255, 255,
255, 255, 255, 255, 255, 0, 0, 0, 0, 255, 255, 255, 255, 255, 255, 255, 255, 255, 0, 0, 0, 0, 255, 255, 255, 255,
255, 255, 255, 255, 255, 255, 0, 0, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 0, 0, 255, 255, 255, 255, 255,
255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255
};
Mat big_image = imread(cvtest::TS::ptr()->get_data_path() + "mser/puzzle.png", 0);
Mat small_image(14, 26, CV_8U, buf);
static const int thresharr[] = { 0, 70, 120, 180, 255 };
const int kDelta = 5;
Ptr<MSER> mserExtractor = MSER::create( kDelta );
vector<vector<Point> > msers;
vector<Rect> boxes;
int CV_MserTest::CompareBoxes(const vector<CvBox2D>& boxes1,const vector<CvBox2D>& boxes2, float max_rel_diff)
{
if (boxes1.size() != boxes2.size())
return 0;
RNG& rng = theRNG();
for (int i=0; i<(int)boxes1.size();i++)
for( int i = 0; i < 30; i++ )
{
float rel_diff;
if (!((boxes1[i].angle == 0.0f) && (abs(boxes2[i].angle) < max_rel_diff)))
{
float angle_diff = (float)fmod(boxes1[i].angle - boxes2[i].angle, 180);
// for angular correctness, it makes no sense to use a "relative" error.
// a 1-degree error around 5 degrees is equally bas as around 250 degrees.
// in correct cases, angle_diff can now be a bit above 0 or a bit below 180
if (angle_diff > 90.0f)
{
angle_diff -= 180.0f;
}
rel_diff = (float)fabs(angle_diff);
if (rel_diff > max_rel_diff)
return i;
}
if (!((boxes1[i].center.x == 0.0f) && (abs(boxes2[i].center.x) < max_rel_diff)))
{
rel_diff = abs(boxes1[i].center.x-boxes2[i].center.x)/abs(boxes1[i].center.x);
if (rel_diff > max_rel_diff)
return i;
}
if (!((boxes1[i].center.y == 0.0f) && (abs(boxes2[i].center.y) < max_rel_diff)))
{
rel_diff = abs(boxes1[i].center.y-boxes2[i].center.y)/abs(boxes1[i].center.y);
if (rel_diff > max_rel_diff)
return i;
}
if (!((boxes1[i].size.width == 0.0f) && (abs(boxes2[i].size.width) < max_rel_diff)))
{
rel_diff = abs(boxes1[i].size.width-boxes2[i].size.width)/abs(boxes1[i].size.width);
if (rel_diff > max_rel_diff)
return i;
}
bool use_big_image = rng.uniform(0, 7) != 0;
bool invert = rng.uniform(0, 2) != 0;
bool binarize = use_big_image ? rng.uniform(0, 5) != 0 : false;
bool blur = rng.uniform(0, 2) != 0;
int thresh = thresharr[rng.uniform(0, 5)];
if (!((boxes1[i].size.height == 0.0f) && (abs(boxes2[i].size.height) < max_rel_diff)))
/*if( i == 0 )
{
rel_diff = abs(boxes1[i].size.height-boxes2[i].size.height)/abs(boxes1[i].size.height);
if (rel_diff > max_rel_diff)
return i;
}
}
use_big_image = true;
invert = binarize = blur = false;
}*/
return -1;
}
const Mat& src0 = use_big_image ? big_image : small_image;
Mat src = src0.clone();
void CV_MserTest::run(int)
{
string image_path = string(ts->get_data_path()) + "mser/puzzle.png";
int kMinArea = use_big_image ? 256 : 10;
int kMaxArea = (int)src.total()/4;
Mat img = imread( image_path );
if (img.empty())
{
ts->printf( cvtest::TS::LOG, "Unable to open image mser/puzzle.png\n");
ts->set_failed_test_info(cvtest::TS::FAIL_MISSING_TEST_DATA);
return;
}
mserExtractor->setMinArea(kMinArea);
mserExtractor->setMaxArea(kMaxArea);
Mat yuv;
cvtColor(img, yuv, COLOR_BGR2YCrCb);
vector<vector<Point> > msers;
MSER()(yuv, msers);
if( invert )
bitwise_not(src, src);
if( binarize )
threshold(src, src, thresh, 255, THRESH_BINARY);
if( blur )
GaussianBlur(src, src, Size(5, 5), 1.5, 1.5);
vector<CvBox2D> boxes;
vector<CvBox2D> boxes_orig;
for ( size_t i = 0; i < msers.size(); i++ )
{
RotatedRect box = fitEllipse(msers[i]);
box.angle=(float)CV_PI/2-box.angle;
boxes.push_back(box);
}
int minRegs = use_big_image ? 10 : 2;
int maxRegs = use_big_image ? 1000 : 10;
if( binarize && (thresh == 0 || thresh == 255) )
minRegs = maxRegs = 0;
string boxes_path = string(ts->get_data_path()) + "mser/boxes.txt";
string calc_boxes_path = string(ts->get_data_path()) + "mser/boxes.calc.txt";
mserExtractor->detectRegions( src, msers, boxes );
int nmsers = (int)msers.size();
ASSERT_EQ(nmsers, (int)boxes.size());
if (!LoadBoxes(boxes_path.c_str(),boxes_orig))
{
SaveBoxes(boxes_path.c_str(),boxes);
ts->printf( cvtest::TS::LOG, "Unable to open data file mser/boxes.txt\n");
ts->set_failed_test_info(cvtest::TS::FAIL_MISSING_TEST_DATA);
return;
}
if( maxRegs < nmsers || minRegs > nmsers )
{
printf("%d. minArea=%d, maxArea=%d, nmsers=%d, minRegs=%d, maxRegs=%d, "
"image=%s, invert=%d, binarize=%d, thresh=%d, blur=%d\n",
i, kMinArea, kMaxArea, nmsers, minRegs, maxRegs, use_big_image ? "big" : "small",
(int)invert, (int)binarize, thresh, (int)blur);
#if defined RENDER_MSERS && RENDER_MSERS
Mat image;
imshow("source", src);
renderMSERs(src, image, msers);
imshow("result", image);
waitKey();
#endif
}
const float dissimularity = 0.01f;
int n_box = CompareBoxes(boxes_orig,boxes,dissimularity);
if (n_box < 0)
{
ts->set_failed_test_info(cvtest::TS::OK);
}
else
{
SaveBoxes(calc_boxes_path.c_str(), boxes);
ts->set_failed_test_info(cvtest::TS::FAIL_BAD_ACCURACY);
ts->printf( cvtest::TS::LOG, "Incorrect correspondence in box %d\n",n_box);
ASSERT_LE(minRegs, nmsers);
ASSERT_GE(maxRegs, nmsers);
}
}
TEST(Features2d_MSER, DISABLED_regression) { CV_MserTest test; test.safe_run(); }
#endif
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment