Merge pull request #1283 from melody-rain:new_mog_mog2

modules/ocl/include/opencv2/ocl/ocl.hpp

@@ -1698,6 +1698,155 @@ namespace cv
         //    keys   = {1,    2,   3}   (CV_8UC1)
         //    values = {6,2, 10,5, 4,3} (CV_8UC2)
         void CV_EXPORTS sortByKey(oclMat& keys, oclMat& values, int method, bool isGreaterThan = false);
+        /*!Base class for MOG and MOG2!*/
+        class CV_EXPORTS BackgroundSubtractor
+        {
+        public:
+            //! the virtual destructor
+            virtual ~BackgroundSubtractor();
+            //! the update operator that takes the next video frame and returns the current foreground mask as 8-bit binary image.
+            virtual void operator()(const oclMat& image, oclMat& fgmask, float learningRate);
+
+            //! computes a background image
+            virtual void getBackgroundImage(oclMat& backgroundImage) const = 0;
+        };
+                /*!
+        Gaussian Mixture-based Backbround/Foreground Segmentation Algorithm
+
+        The class implements the following algorithm:
+        "An improved adaptive background mixture model for real-time tracking with shadow detection"
+        P. KadewTraKuPong and R. Bowden,
+        Proc. 2nd European Workshp on Advanced Video-Based Surveillance Systems, 2001."
+        http://personal.ee.surrey.ac.uk/Personal/R.Bowden/publications/avbs01/avbs01.pdf
+        */
+        class CV_EXPORTS MOG: public cv::ocl::BackgroundSubtractor
+        {
+        public:
+            //! the default constructor
+            MOG(int nmixtures = -1);
+
+            //! re-initiaization method
+            void initialize(Size frameSize, int frameType);
+
+            //! the update operator
+            void operator()(const oclMat& frame, oclMat& fgmask, float learningRate = 0.f);
+
+            //! computes a background image which are the mean of all background gaussians
+            void getBackgroundImage(oclMat& backgroundImage) const;
+
+            //! releases all inner buffers
+            void release();
+
+            int history;
+            float varThreshold;
+            float backgroundRatio;
+            float noiseSigma;
+
+        private:
+            int nmixtures_;
+
+            Size frameSize_;
+            int frameType_;
+            int nframes_;
+
+            oclMat weight_;
+            oclMat sortKey_;
+            oclMat mean_;
+            oclMat var_;
+        };
+
+        /*!
+        The class implements the following algorithm:
+        "Improved adaptive Gausian mixture model for background subtraction"
+        Z.Zivkovic
+        International Conference Pattern Recognition, UK, August, 2004.
+        http://www.zoranz.net/Publications/zivkovic2004ICPR.pdf
+        */
+        class CV_EXPORTS MOG2: public cv::ocl::BackgroundSubtractor
+        {
+        public:
+            //! the default constructor
+            MOG2(int nmixtures = -1);
+
+            //! re-initiaization method
+            void initialize(Size frameSize, int frameType);
+
+            //! the update operator
+            void operator()(const oclMat& frame, oclMat& fgmask, float learningRate = -1.0f);
+
+            //! computes a background image which are the mean of all background gaussians
+            void getBackgroundImage(oclMat& backgroundImage) const;
+
+            //! releases all inner buffers
+            void release();
+
+            // parameters
+            // you should call initialize after parameters changes
+
+            int history;
+
+            //! here it is the maximum allowed number of mixture components.
+            //! Actual number is determined dynamically per pixel
+            float varThreshold;
+            // threshold on the squared Mahalanobis distance to decide if it is well described
+            // by the background model or not. Related to Cthr from the paper.
+            // This does not influence the update of the background. A typical value could be 4 sigma
+            // and that is varThreshold=4*4=16; Corresponds to Tb in the paper.
+
+            /////////////////////////
+            // less important parameters - things you might change but be carefull
+            ////////////////////////
+
+            float backgroundRatio;
+            // corresponds to fTB=1-cf from the paper
+            // TB - threshold when the component becomes significant enough to be included into
+            // the background model. It is the TB=1-cf from the paper. So I use cf=0.1 => TB=0.
+            // For alpha=0.001 it means that the mode should exist for approximately 105 frames before
+            // it is considered foreground
+            // float noiseSigma;
+            float varThresholdGen;
+
+            //correspondts to Tg - threshold on the squared Mahalan. dist. to decide
+            //when a sample is close to the existing components. If it is not close
+            //to any a new component will be generated. I use 3 sigma => Tg=3*3=9.
+            //Smaller Tg leads to more generated components and higher Tg might make
+            //lead to small number of components but they can grow too large
+            float fVarInit;
+            float fVarMin;
+            float fVarMax;
+
+            //initial variance  for the newly generated components.
+            //It will will influence the speed of adaptation. A good guess should be made.
+            //A simple way is to estimate the typical standard deviation from the images.
+            //I used here 10 as a reasonable value
+            // min and max can be used to further control the variance
+            float fCT; //CT - complexity reduction prior
+            //this is related to the number of samples needed to accept that a component
+            //actually exists. We use CT=0.05 of all the samples. By setting CT=0 you get
+            //the standard Stauffer&Grimson algorithm (maybe not exact but very similar)
+
+            //shadow detection parameters
+            bool bShadowDetection; //default 1 - do shadow detection
+            unsigned char nShadowDetection; //do shadow detection - insert this value as the detection result - 127 default value
+            float fTau;
+            // Tau - shadow threshold. The shadow is detected if the pixel is darker
+            //version of the background. Tau is a threshold on how much darker the shadow can be.
+            //Tau= 0.5 means that if pixel is more than 2 times darker then it is not shadow
+            //See: Prati,Mikic,Trivedi,Cucchiarra,"Detecting Moving Shadows...",IEEE PAMI,2003.
+
+        private:
+            int nmixtures_;
+
+            Size frameSize_;
+            int frameType_;
+            int nframes_;
+
+            oclMat weight_;
+            oclMat variance_;
+            oclMat mean_;
+
+            oclMat bgmodelUsedModes_; //keep track of number of modes per pixel
+        };
     }
 }
 #if defined _MSC_VER && _MSC_VER >= 1200

modules/ocl/test/test_bgfg.cpp

+/*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) 2010-2013, Multicoreware, Inc., all rights reserved.
+// Copyright (C) 2010-2013, Advanced Micro Devices, Inc., all rights reserved.
+// Third party copyrights are property of their respective owners.
+//
+// @Authors
+//    Jin Ma, jin@multicorewareinc.com
+//
+// 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 oclMaterials 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 "test_precomp.hpp"
+
+#ifdef HAVE_OPENCL
+
+using namespace cv;
+using namespace cv::ocl;
+using namespace cvtest;
+using namespace testing;
+using namespace std;
+
+extern string workdir;
+//////////////////////////////////////////////////////
+// MOG
+
+namespace
+{
+    IMPLEMENT_PARAM_CLASS(UseGray, bool)
+    IMPLEMENT_PARAM_CLASS(LearningRate, double)
+}
+
+PARAM_TEST_CASE(mog, UseGray, LearningRate, bool)
+{
+    bool useGray;
+    double learningRate;
+    bool useRoi;
+
+    virtual void SetUp()
+    {
+        useGray = GET_PARAM(0);
+
+        learningRate = GET_PARAM(1);
+
+        useRoi = GET_PARAM(2);
+    }
+};
+
+TEST_P(mog, Update)
+{
+    std::string inputFile = string(cvtest::TS::ptr()->get_data_path()) + "gpu/768x576.avi";
+    cv::VideoCapture cap(inputFile);
+    ASSERT_TRUE(cap.isOpened());
+
+    cv::Mat frame;
+    cap >> frame;
+    ASSERT_FALSE(frame.empty());
+
+    cv::ocl::MOG mog;
+    cv::ocl::oclMat foreground = createMat_ocl(frame.size(), CV_8UC1, useRoi);
+
+    cv::BackgroundSubtractorMOG mog_gold;
+    cv::Mat foreground_gold;
+
+    for (int i = 0; i < 10; ++i)
+    {
+        cap >> frame;
+        ASSERT_FALSE(frame.empty());
+
+        if (useGray)
+        {
+            cv::Mat temp;
+            cv::cvtColor(frame, temp, cv::COLOR_BGR2GRAY);
+            cv::swap(temp, frame);
+        }
+
+        mog(loadMat_ocl(frame, useRoi), foreground, (float)learningRate);
+
+        mog_gold(frame, foreground_gold, learningRate);
+
+        EXPECT_MAT_NEAR(foreground_gold, foreground, 0.0);
+    }
+}
+INSTANTIATE_TEST_CASE_P(OCL_Video, mog, testing::Combine(
+    testing::Values(UseGray(false), UseGray(true)),
+    testing::Values(LearningRate(0.0), LearningRate(0.01)),
+    Values(true, false)));
+
+//////////////////////////////////////////////////////
+// MOG2
+
+namespace
+{
+    IMPLEMENT_PARAM_CLASS(DetectShadow, bool)
+}
+
+PARAM_TEST_CASE(mog2, UseGray, DetectShadow, bool)
+{
+    bool useGray;
+    bool detectShadow;
+    bool useRoi;
+    virtual void SetUp()
+    {
+        useGray = GET_PARAM(0);
+        detectShadow = GET_PARAM(1);
+        useRoi = GET_PARAM(2);
+    }
+};
+
+TEST_P(mog2, Update)
+{
+    std::string inputFile = string(cvtest::TS::ptr()->get_data_path()) + "gpu/768x576.avi";
+    cv::VideoCapture cap(inputFile);
+    ASSERT_TRUE(cap.isOpened());
+
+    cv::Mat frame;
+    cap >> frame;
+    ASSERT_FALSE(frame.empty());
+
+    cv::ocl::MOG2 mog2;
+    mog2.bShadowDetection = detectShadow;
+    cv::ocl::oclMat foreground = createMat_ocl(frame.size(), CV_8UC1, useRoi);
+
+    cv::BackgroundSubtractorMOG2 mog2_gold;
+    mog2_gold.set("detectShadows", detectShadow);
+    cv::Mat foreground_gold;
+
+    for (int i = 0; i < 10; ++i)
+    {
+        cap >> frame;
+        ASSERT_FALSE(frame.empty());
+
+        if (useGray)
+        {
+            cv::Mat temp;
+            cv::cvtColor(frame, temp, cv::COLOR_BGR2GRAY);
+            cv::swap(temp, frame);
+        }
+
+        mog2(loadMat_ocl(frame, useRoi), foreground);
+
+        mog2_gold(frame, foreground_gold);
+
+        if (detectShadow)
+        {
+
+            EXPECT_MAT_SIMILAR(foreground_gold, foreground, 1e-2);
+        }
+        else
+        {
+            EXPECT_MAT_NEAR(foreground_gold, foreground, 0);
+        }
+    }
+}
+
+TEST_P(mog2, getBackgroundImage)
+{
+    if (useGray)
+        return;
+
+    std::string inputFile = string(cvtest::TS::ptr()->get_data_path()) + "video/768x576.avi";
+    cv::VideoCapture cap(inputFile);
+    ASSERT_TRUE(cap.isOpened());
+
+    cv::Mat frame;
+
+    cv::ocl::MOG2 mog2;
+    mog2.bShadowDetection = detectShadow;
+    cv::ocl::oclMat foreground;
+
+    cv::BackgroundSubtractorMOG2 mog2_gold;
+    mog2_gold.set("detectShadows", detectShadow);
+    cv::Mat foreground_gold;
+
+    for (int i = 0; i < 10; ++i)
+    {
+        cap >> frame;
+        ASSERT_FALSE(frame.empty());
+
+        mog2(loadMat_ocl(frame, useRoi), foreground);
+
+        mog2_gold(frame, foreground_gold);
+    }
+
+    cv::ocl::oclMat background = createMat_ocl(frame.size(), frame.type(), useRoi);
+    mog2.getBackgroundImage(background);
+
+    cv::Mat background_gold;
+    mog2_gold.getBackgroundImage(background_gold);
+
+    EXPECT_MAT_NEAR(background_gold, background, 1.0);
+}
+
+INSTANTIATE_TEST_CASE_P(OCL_Video, mog2, testing::Combine(
+    testing::Values(UseGray(true), UseGray(false)),
+    testing::Values(DetectShadow(true), DetectShadow(false)),
+    Values(true, false)));
+
+#endif
\ No newline at end of file

modules/ocl/test/test_optflow.cpp

@@ -146,10 +146,10 @@ PARAM_TEST_CASE(TVL1, bool)
 
 TEST_P(TVL1, Accuracy)
 {
-    cv::Mat frame0 = readImage("gpu/opticalflow/rubberwhale1.png", cv::IMREAD_GRAYSCALE);
+    cv::Mat frame0 = readImage("F:/mcw/opencv/opencv/samples/gpu/rubberwhale1.png", cv::IMREAD_GRAYSCALE);
     ASSERT_FALSE(frame0.empty());
 
-    cv::Mat frame1 = readImage("gpu/opticalflow/rubberwhale2.png", cv::IMREAD_GRAYSCALE);
+    cv::Mat frame1 = readImage("../../../opencv/samples/gpu/rubberwhale2.png", cv::IMREAD_GRAYSCALE);
     ASSERT_FALSE(frame1.empty());
 
     cv::ocl::OpticalFlowDual_TVL1_OCL d_alg;
@@ -168,7 +168,7 @@ TEST_P(TVL1, Accuracy)
     EXPECT_MAT_SIMILAR(gold[0], d_flowx, 3e-3);
     EXPECT_MAT_SIMILAR(gold[1], d_flowy, 3e-3);
 }
-INSTANTIATE_TEST_CASE_P(OCL_Video, TVL1, Values(true, false));
+INSTANTIATE_TEST_CASE_P(OCL_Video, TVL1, Values(false, true));
 
 
 /////////////////////////////////////////////////////////////////////////////////////////////////

modules/ocl/test/utility.cpp

@@ -100,6 +100,44 @@ Mat randomMat(Size size, int type, double minVal, double maxVal)
     return randomMat(TS::ptr()->get_rng(), size, type, minVal, maxVal, false);
 }
 
+cv::ocl::oclMat createMat_ocl(Size size, int type, bool useRoi)
+{
+    Size size0 = size;
+
+    if (useRoi)
+    {
+        size0.width += randomInt(5, 15);
+        size0.height += randomInt(5, 15);
+    }
+
+    cv::ocl::oclMat d_m(size0, type);
+
+    if (size0 != size)
+        d_m = d_m(Rect((size0.width - size.width) / 2, (size0.height - size.height) / 2, size.width, size.height));
+
+    return d_m;
+}
+
+cv::ocl::oclMat loadMat_ocl(const Mat& m, bool useRoi)
+{
+    CV_Assert(m.type() == CV_8UC1 || m.type() == CV_8UC3);
+    cv::ocl::oclMat d_m;
+    d_m = createMat_ocl(m.size(), m.type(), useRoi);
+
+    Size ls;
+    Point pt;
+
+    d_m.locateROI(ls, pt);
+
+    Rect roi(pt.x, pt.y, d_m.size().width, d_m.size().height);
+
+    cv::ocl::oclMat m_ocl(m);
+
+    cv::ocl::oclMat d_m_roi(d_m, roi);
+
+    m_ocl.copyTo(d_m);
+    return d_m;
+}
 /*
 void showDiff(InputArray gold_, InputArray actual_, double eps)
 {

modules/ocl/test/utility.hpp

@@ -70,6 +70,9 @@ double checkNorm(const cv::Mat &m);
 double checkNorm(const cv::Mat &m1, const cv::Mat &m2);
 double checkSimilarity(const cv::Mat &m1, const cv::Mat &m2);
 
+//oclMat create
+cv::ocl::oclMat createMat_ocl(cv::Size size, int type, bool useRoi = false);
+cv::ocl::oclMat loadMat_ocl(const cv::Mat& m, bool useRoi = false);
 #define EXPECT_MAT_NORM(mat, eps) \
 { \
     EXPECT_LE(checkNorm(cv::Mat(mat)), eps) \

samples/ocl/bgfg_segm.cpp

+#include <iostream>
+#include <string>
+
+#include "opencv2/core/core.hpp"
+#include "opencv2/ocl/ocl.hpp"
+#include "opencv2/highgui/highgui.hpp"
+
+using namespace std;
+using namespace cv;
+using namespace cv::ocl;
+
+#define M_MOG  1
+#define M_MOG2 2
+
+int main(int argc, const char** argv)
+{
+
+    cv::CommandLineParser cmd(argc, argv,
+        "{ c | camera | false       | use camera }"
+        "{ f | file   | 768x576.avi | input video file }"
+        "{ m | method | mog         | method (mog, mog2) }"
+        "{ h | help   | false       | print help message }");
+
+    if (cmd.get<bool>("help"))
+    {
+        cout << "Usage : bgfg_segm [options]" << endl;
+        cout << "Avaible options:" << endl;
+        cmd.printParams();
+        return 0;
+    }
+
+    bool useCamera = cmd.get<bool>("camera");
+    string file = cmd.get<string>("file");
+    string method = cmd.get<string>("method");
+
+    if (method != "mog" && method != "mog2")
+    {
+        cerr << "Incorrect method" << endl;
+        return -1;
+    }
+
+    int m = method == "mog" ? M_MOG : M_MOG2;
+
+    VideoCapture cap;
+
+    if (useCamera)
+        cap.open(0);
+    else
+        cap.open(file);
+
+    if (!cap.isOpened())
+    {
+        cerr << "can not open camera or video file" << endl;
+        return -1;
+    }
+
+    std::vector<cv::ocl::Info>info;
+    cv::ocl::getDevice(info);
+
+    Mat frame;
+    cap >> frame;
+
+    oclMat d_frame(frame);
+
+    cv::ocl::MOG mog;
+    cv::ocl::MOG2 mog2;
+
+    oclMat d_fgmask;
+    oclMat d_fgimg;
+    oclMat d_bgimg;
+
+    d_fgimg.create(d_frame.size(), d_frame.type());
+
+    Mat fgmask;
+    Mat fgimg;
+    Mat bgimg;
+
+    switch (m)
+    {
+    case M_MOG:
+        mog(d_frame, d_fgmask, 0.01f);
+        break;
+
+    case M_MOG2:
+        mog2(d_frame, d_fgmask);
+        break;
+    }
+
+    for(;;)
+    {
+        cap >> frame;
+        if (frame.empty())
+            break;
+        d_frame.upload(frame);
+
+        int64 start = cv::getTickCount();
+
+        //update the model
+        switch (m)
+        {
+        case M_MOG:
+            mog(d_frame, d_fgmask, 0.01f);
+            mog.getBackgroundImage(d_bgimg);
+            break;
+
+        case M_MOG2:
+            mog2(d_frame, d_fgmask);
+            mog2.getBackgroundImage(d_bgimg);
+            break;
+        }
+
+        double fps = cv::getTickFrequency() / (cv::getTickCount() - start);
+        std::cout << "FPS : " << fps << std::endl;
+
+        d_fgimg.setTo(Scalar::all(0));
+        d_frame.copyTo(d_fgimg, d_fgmask);
+
+        d_fgmask.download(fgmask);
+        d_fgimg.download(fgimg);
+        if (!d_bgimg.empty())
+            d_bgimg.download(bgimg);
+
+        imshow("image", frame);
+        imshow("foreground mask", fgmask);
+        imshow("foreground image", fgimg);
+        if (!bgimg.empty())
+            imshow("mean background image", bgimg);
+
+        int key = waitKey(30);
+        if (key == 27)
+            break;
+    }
+
+    return 0;
+}