opencv/modules/imgproc/src/featureselect.cpp
Matthew Self 72672c293f Make goodFeaturesToTrack() return deterministic results
When using OCL, the results of goodFeaturesToTrack() vary slightly from
run to run. This appears to be because the order of the results from
the findCorners kernel depends on thread execution and the sorting
function that is used at the end to rank the features only enforces are
partial sort order.

This does not materially impact the quality of the results, but it
makes it hard to build regression tests and generally introduces noise
into the system that should be avoided.

An easy fix is to change the sort function to enforce a total sort on
the features, even in cases where the match quality is exactly the same
for two features.
2016-07-25 21:17:54 -07:00

441 lines
15 KiB
C++

/*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.
//
//
// Intel License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000, Intel Corporation, 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 "opencl_kernels_imgproc.hpp"
#include <cstdio>
#include <vector>
#include <iostream>
#include <functional>
namespace cv
{
struct greaterThanPtr :
public std::binary_function<const float *, const float *, bool>
{
bool operator () (const float * a, const float * b) const
// Ensure a fully deterministic result of the sort
{ return (*a > *b) ? true : (*a < *b) ? false : (a > b); }
};
#ifdef HAVE_OPENCL
struct Corner
{
float val;
short y;
short x;
bool operator < (const Corner & c) const
// Ensure a fully deterministic result of the sort
{ return (val > c.val) ? true : (val < c.val) ? false : (y > c.y) ? true : (y < c.y) ? false : (x > c.x); }
};
static bool ocl_goodFeaturesToTrack( InputArray _image, OutputArray _corners,
int maxCorners, double qualityLevel, double minDistance,
InputArray _mask, int blockSize,
bool useHarrisDetector, double harrisK )
{
UMat eig, maxEigenValue;
if( useHarrisDetector )
cornerHarris( _image, eig, blockSize, 3, harrisK );
else
cornerMinEigenVal( _image, eig, blockSize, 3 );
Size imgsize = _image.size();
size_t total, i, j, ncorners = 0, possibleCornersCount =
std::max(1024, static_cast<int>(imgsize.area() * 0.1));
bool haveMask = !_mask.empty();
UMat corners_buffer(1, (int)possibleCornersCount + 1, CV_32FC2);
CV_Assert(sizeof(Corner) == corners_buffer.elemSize());
Mat tmpCorners;
// find threshold
{
CV_Assert(eig.type() == CV_32FC1);
int dbsize = ocl::Device::getDefault().maxComputeUnits();
size_t wgs = ocl::Device::getDefault().maxWorkGroupSize();
int wgs2_aligned = 1;
while (wgs2_aligned < (int)wgs)
wgs2_aligned <<= 1;
wgs2_aligned >>= 1;
ocl::Kernel k("maxEigenVal", ocl::imgproc::gftt_oclsrc,
format("-D OP_MAX_EIGEN_VAL -D WGS=%d -D groupnum=%d -D WGS2_ALIGNED=%d%s",
(int)wgs, dbsize, wgs2_aligned, haveMask ? " -D HAVE_MASK" : ""));
if (k.empty())
return false;
UMat mask = _mask.getUMat();
maxEigenValue.create(1, dbsize, CV_32FC1);
ocl::KernelArg eigarg = ocl::KernelArg::ReadOnlyNoSize(eig),
dbarg = ocl::KernelArg::PtrWriteOnly(maxEigenValue),
maskarg = ocl::KernelArg::ReadOnlyNoSize(mask),
cornersarg = ocl::KernelArg::PtrWriteOnly(corners_buffer);
if (haveMask)
k.args(eigarg, eig.cols, (int)eig.total(), dbarg, maskarg);
else
k.args(eigarg, eig.cols, (int)eig.total(), dbarg);
size_t globalsize = dbsize * wgs;
if (!k.run(1, &globalsize, &wgs, false))
return false;
ocl::Kernel k2("maxEigenValTask", ocl::imgproc::gftt_oclsrc,
format("-D OP_MAX_EIGEN_VAL -D WGS=%d -D WGS2_ALIGNED=%d -D groupnum=%d",
wgs, wgs2_aligned, dbsize));
if (k2.empty())
return false;
k2.args(dbarg, (float)qualityLevel, cornersarg);
if (!k2.runTask(false))
return false;
}
// collect list of pointers to features - put them into temporary image
{
ocl::Kernel k("findCorners", ocl::imgproc::gftt_oclsrc,
format("-D OP_FIND_CORNERS%s", haveMask ? " -D HAVE_MASK" : ""));
if (k.empty())
return false;
ocl::KernelArg eigarg = ocl::KernelArg::ReadOnlyNoSize(eig),
cornersarg = ocl::KernelArg::PtrWriteOnly(corners_buffer),
thresholdarg = ocl::KernelArg::PtrReadOnly(maxEigenValue);
if (!haveMask)
k.args(eigarg, cornersarg, eig.rows - 2, eig.cols - 2, thresholdarg,
(int)possibleCornersCount);
else
{
UMat mask = _mask.getUMat();
k.args(eigarg, ocl::KernelArg::ReadOnlyNoSize(mask),
cornersarg, eig.rows - 2, eig.cols - 2,
thresholdarg, (int)possibleCornersCount);
}
size_t globalsize[2] = { (size_t)eig.cols - 2, (size_t)eig.rows - 2 };
if (!k.run(2, globalsize, NULL, false))
return false;
tmpCorners = corners_buffer.getMat(ACCESS_RW);
total = std::min<size_t>(tmpCorners.at<Vec2i>(0, 0)[0], possibleCornersCount);
if (total == 0)
{
_corners.release();
return true;
}
}
Corner* corner_ptr = tmpCorners.ptr<Corner>() + 1;
std::sort(corner_ptr, corner_ptr + total);
std::vector<Point2f> corners;
corners.reserve(total);
if (minDistance >= 1)
{
// Partition the image into larger grids
int w = imgsize.width, h = imgsize.height;
const int cell_size = cvRound(minDistance);
const int grid_width = (w + cell_size - 1) / cell_size;
const int grid_height = (h + cell_size - 1) / cell_size;
std::vector<std::vector<Point2f> > grid(grid_width*grid_height);
minDistance *= minDistance;
for( i = 0; i < total; i++ )
{
const Corner & c = corner_ptr[i];
bool good = true;
int x_cell = c.x / cell_size;
int y_cell = c.y / cell_size;
int x1 = x_cell - 1;
int y1 = y_cell - 1;
int x2 = x_cell + 1;
int y2 = y_cell + 1;
// boundary check
x1 = std::max(0, x1);
y1 = std::max(0, y1);
x2 = std::min(grid_width - 1, x2);
y2 = std::min(grid_height - 1, y2);
for( int yy = y1; yy <= y2; yy++ )
for( int xx = x1; xx <= x2; xx++ )
{
std::vector<Point2f> &m = grid[yy * grid_width + xx];
if( m.size() )
{
for(j = 0; j < m.size(); j++)
{
float dx = c.x - m[j].x;
float dy = c.y - m[j].y;
if( dx*dx + dy*dy < minDistance )
{
good = false;
goto break_out;
}
}
}
}
break_out:
if (good)
{
grid[y_cell*grid_width + x_cell].push_back(Point2f((float)c.x, (float)c.y));
corners.push_back(Point2f((float)c.x, (float)c.y));
++ncorners;
if( maxCorners > 0 && (int)ncorners == maxCorners )
break;
}
}
}
else
{
for( i = 0; i < total; i++ )
{
const Corner & c = corner_ptr[i];
corners.push_back(Point2f((float)c.x, (float)c.y));
++ncorners;
if( maxCorners > 0 && (int)ncorners == maxCorners )
break;
}
}
Mat(corners).convertTo(_corners, _corners.fixedType() ? _corners.type() : CV_32F);
return true;
}
#endif
}
void cv::goodFeaturesToTrack( InputArray _image, OutputArray _corners,
int maxCorners, double qualityLevel, double minDistance,
InputArray _mask, int blockSize,
bool useHarrisDetector, double harrisK )
{
CV_Assert( qualityLevel > 0 && minDistance >= 0 && maxCorners >= 0 );
CV_Assert( _mask.empty() || (_mask.type() == CV_8UC1 && _mask.sameSize(_image)) );
CV_OCL_RUN(_image.dims() <= 2 && _image.isUMat(),
ocl_goodFeaturesToTrack(_image, _corners, maxCorners, qualityLevel, minDistance,
_mask, blockSize, useHarrisDetector, harrisK))
Mat image = _image.getMat(), eig, tmp;
if (image.empty())
{
_corners.release();
return;
}
if( useHarrisDetector )
cornerHarris( image, eig, blockSize, 3, harrisK );
else
cornerMinEigenVal( image, eig, blockSize, 3 );
double maxVal = 0;
minMaxLoc( eig, 0, &maxVal, 0, 0, _mask );
threshold( eig, eig, maxVal*qualityLevel, 0, THRESH_TOZERO );
dilate( eig, tmp, Mat());
Size imgsize = image.size();
std::vector<const float*> tmpCorners;
// collect list of pointers to features - put them into temporary image
Mat mask = _mask.getMat();
for( int y = 1; y < imgsize.height - 1; y++ )
{
const float* eig_data = (const float*)eig.ptr(y);
const float* tmp_data = (const float*)tmp.ptr(y);
const uchar* mask_data = mask.data ? mask.ptr(y) : 0;
for( int x = 1; x < imgsize.width - 1; x++ )
{
float val = eig_data[x];
if( val != 0 && val == tmp_data[x] && (!mask_data || mask_data[x]) )
tmpCorners.push_back(eig_data + x);
}
}
std::vector<Point2f> corners;
size_t i, j, total = tmpCorners.size(), ncorners = 0;
if (total == 0)
{
_corners.release();
return;
}
std::sort( tmpCorners.begin(), tmpCorners.end(), greaterThanPtr() );
if (minDistance >= 1)
{
// Partition the image into larger grids
int w = image.cols;
int h = image.rows;
const int cell_size = cvRound(minDistance);
const int grid_width = (w + cell_size - 1) / cell_size;
const int grid_height = (h + cell_size - 1) / cell_size;
std::vector<std::vector<Point2f> > grid(grid_width*grid_height);
minDistance *= minDistance;
for( i = 0; i < total; i++ )
{
int ofs = (int)((const uchar*)tmpCorners[i] - eig.ptr());
int y = (int)(ofs / eig.step);
int x = (int)((ofs - y*eig.step)/sizeof(float));
bool good = true;
int x_cell = x / cell_size;
int y_cell = y / cell_size;
int x1 = x_cell - 1;
int y1 = y_cell - 1;
int x2 = x_cell + 1;
int y2 = y_cell + 1;
// boundary check
x1 = std::max(0, x1);
y1 = std::max(0, y1);
x2 = std::min(grid_width-1, x2);
y2 = std::min(grid_height-1, y2);
for( int yy = y1; yy <= y2; yy++ )
{
for( int xx = x1; xx <= x2; xx++ )
{
std::vector <Point2f> &m = grid[yy*grid_width + xx];
if( m.size() )
{
for(j = 0; j < m.size(); j++)
{
float dx = x - m[j].x;
float dy = y - m[j].y;
if( dx*dx + dy*dy < minDistance )
{
good = false;
goto break_out;
}
}
}
}
}
break_out:
if (good)
{
grid[y_cell*grid_width + x_cell].push_back(Point2f((float)x, (float)y));
corners.push_back(Point2f((float)x, (float)y));
++ncorners;
if( maxCorners > 0 && (int)ncorners == maxCorners )
break;
}
}
}
else
{
for( i = 0; i < total; i++ )
{
int ofs = (int)((const uchar*)tmpCorners[i] - eig.ptr());
int y = (int)(ofs / eig.step);
int x = (int)((ofs - y*eig.step)/sizeof(float));
corners.push_back(Point2f((float)x, (float)y));
++ncorners;
if( maxCorners > 0 && (int)ncorners == maxCorners )
break;
}
}
Mat(corners).convertTo(_corners, _corners.fixedType() ? _corners.type() : CV_32F);
}
CV_IMPL void
cvGoodFeaturesToTrack( const void* _image, void*, void*,
CvPoint2D32f* _corners, int *_corner_count,
double quality_level, double min_distance,
const void* _maskImage, int block_size,
int use_harris, double harris_k )
{
cv::Mat image = cv::cvarrToMat(_image), mask;
std::vector<cv::Point2f> corners;
if( _maskImage )
mask = cv::cvarrToMat(_maskImage);
CV_Assert( _corners && _corner_count );
cv::goodFeaturesToTrack( image, corners, *_corner_count, quality_level,
min_distance, mask, block_size, use_harris != 0, harris_k );
size_t i, ncorners = corners.size();
for( i = 0; i < ncorners; i++ )
_corners[i] = corners[i];
*_corner_count = (int)ncorners;
}
/* End of file. */