Refactored SVMSGD class

modules/ml/test/test_mltests2.cpp

@@ -193,6 +193,16 @@ int str_to_boost_type( String& str )
 // 8. rtrees
 // 9. ertrees
 
+int str_to_svmsgd_type( String& str )
+{
+    if ( !str.compare("SGD") )
+        return SVMSGD::SGD;
+    if ( !str.compare("ASGD") )
+        return SVMSGD::ASGD;
+    CV_Error( CV_StsBadArg, "incorrect boost type string" );
+    return -1;
+}
+
 // ---------------------------------- MLBaseTest ---------------------------------------------------
 
 CV_MLBaseTest::CV_MLBaseTest(const char* _modelName)
@@ -248,7 +258,9 @@ void CV_MLBaseTest::run( int )
 {
     string filename = ts->get_data_path();
     filename += get_validation_filename();
+
     validationFS.open( filename, FileStorage::READ );
+
     read_params( *validationFS );
 
     int code = cvtest::TS::OK;
@@ -436,6 +448,21 @@ int CV_MLBaseTest::train( int testCaseIdx )
         model = m;
     }
 
+    else if( modelName == CV_SVMSGD )
+    {
+        String svmsgdTypeStr;
+        modelParamsNode["svmsgdType"] >> svmsgdTypeStr;
+        Ptr<SVMSGD> m = SVMSGD::create();      
+        int type = str_to_svmsgd_type( svmsgdTypeStr );
+        m->setType(type);
+        //m->setType(str_to_svmsgd_type( svmsgdTypeStr ));
+        m->setLambda(modelParamsNode["lambda"]);
+        m->setGamma0(modelParamsNode["gamma0"]);
+        m->setC(modelParamsNode["c"]);
+        m->setTermCriteria(TermCriteria(TermCriteria::COUNT + TermCriteria::EPS, 10000, 0.00001));
+        model = m;
+    }
+
     if( !model.empty() )
         is_trained = model->train(data, 0);
 
@@ -457,7 +484,7 @@ float CV_MLBaseTest::get_test_error( int /*testCaseIdx*/, vector<float> *resp )
     else if( modelName == CV_ANN )
         err = ann_calc_error( model, data, cls_map, type, resp );
     else if( modelName == CV_DTREE || modelName == CV_BOOST || modelName == CV_RTREES ||
-             modelName == CV_SVM || modelName == CV_NBAYES || modelName == CV_KNEAREST )
+             modelName == CV_SVM || modelName == CV_NBAYES || modelName == CV_KNEAREST || modelName == CV_SVMSGD )
         err = model->calcError( data, true, _resp );
     if( !_resp.empty() && resp )
         _resp.convertTo(*resp, CV_32F);
@@ -485,6 +512,8 @@ void CV_MLBaseTest::load( const char* filename )
         model = Algorithm::load<Boost>( filename );
     else if( modelName == CV_RTREES )
         model = Algorithm::load<RTrees>( filename );
+    else if( modelName == CV_SVMSGD )
+        model = Algorithm::load<SVMSGD>( filename );
     else
         CV_Error( CV_StsNotImplemented, "invalid stat model name");
 }

modules/ml/test/test_precomp.hpp

@@ -13,6 +13,7 @@
 #include <map>
 #include "opencv2/ts.hpp"
 #include "opencv2/ml.hpp"
+#include "opencv2/ml/svmsgd.hpp"
 #include "opencv2/core/core_c.h"
 
 #define CV_NBAYES   "nbayes"
@@ -24,6 +25,7 @@
 #define CV_BOOST    "boost"
 #define CV_RTREES   "rtrees"
 #define CV_ERTREES  "ertrees"
+#define CV_SVMSGD   "svmsgd"
 
 enum { CV_TRAIN_ERROR=0, CV_TEST_ERROR=1 };
 
@@ -38,6 +40,7 @@ using cv::ml::ANN_MLP;
 using cv::ml::DTrees;
 using cv::ml::Boost;
 using cv::ml::RTrees;
+using cv::ml::SVMSGD;
 
 class CV_MLBaseTest : public cvtest::BaseTest
 {

modules/ml/test/test_save_load.cpp

@@ -150,12 +150,20 @@ int CV_SLMLTest::validate_test_results( int testCaseIdx )
 
 TEST(ML_NaiveBayes, save_load) { CV_SLMLTest test( CV_NBAYES ); test.safe_run(); }
 TEST(ML_KNearest, save_load) { CV_SLMLTest test( CV_KNEAREST ); test.safe_run(); }
-TEST(ML_SVM, save_load) { CV_SLMLTest test( CV_SVM ); test.safe_run(); }
+TEST(ML_SVM, save_load)
+{
+    CV_SLMLTest test( CV_SVM );
+    test.safe_run();
+}
 TEST(ML_ANN, save_load) { CV_SLMLTest test( CV_ANN ); test.safe_run(); }
 TEST(ML_DTree, save_load) { CV_SLMLTest test( CV_DTREE ); test.safe_run(); }
 TEST(ML_Boost, save_load) { CV_SLMLTest test( CV_BOOST ); test.safe_run(); }
 TEST(ML_RTrees, save_load) { CV_SLMLTest test( CV_RTREES ); test.safe_run(); }
 TEST(DISABLED_ML_ERTrees, save_load) { CV_SLMLTest test( CV_ERTREES ); test.safe_run(); }
+TEST(MV_SVMSGD, save_load){
+    CV_SLMLTest test( CV_SVMSGD );
+    test.safe_run();
+}
 
 class CV_LegacyTest : public cvtest::BaseTest
 {
@@ -201,6 +209,8 @@ protected:
             model = Algorithm::load<SVM>(filename);
         else if (modelName == CV_RTREES)
             model = Algorithm::load<RTrees>(filename);
+        else if (modelName == CV_SVMSGD)
+            model = Algorithm::load<SVMSGD>(filename);
         if (!model)
         {
             code = cvtest::TS::FAIL_INVALID_TEST_DATA;
@@ -260,6 +270,11 @@ TEST(ML_DTree, legacy_load) { CV_LegacyTest test(CV_DTREE, "_abalone.xml;_mushro
 TEST(ML_NBayes, legacy_load) { CV_LegacyTest test(CV_NBAYES, "_waveform.xml"); test.safe_run(); }
 TEST(ML_SVM, legacy_load) { CV_LegacyTest test(CV_SVM, "_poletelecomm.xml;_waveform.xml"); test.safe_run(); }
 TEST(ML_RTrees, legacy_load) { CV_LegacyTest test(CV_RTREES, "_waveform.xml"); test.safe_run(); }
+TEST(ML_SVMSGD, legacy_load)
+{
+    CV_LegacyTest test(CV_SVMSGD, "_waveform.xml");
+    test.safe_run();
+}
 
 /*TEST(ML_SVM, throw_exception_when_save_untrained_model)
 {

modules/ml/test/test_svmsgd.cpp

@@ -0,0 +1,182 @@
+/*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 "test_precomp.hpp"
+#include "opencv2/highgui.hpp"
+
+using namespace cv;
+using namespace cv::ml;
+using cv::ml::SVMSGD;
+using cv::ml::TrainData;
+
+
+
+class CV_SVMSGDTrainTest : public cvtest::BaseTest
+{
+public:
+    CV_SVMSGDTrainTest(Mat _weights, float _shift);
+private:
+    virtual void run( int start_from );
+    float decisionFunction(Mat sample, Mat weights, float shift);
+
+    cv::Ptr<TrainData> data;
+    cv::Mat testSamples;
+    cv::Mat testResponses;
+    static const int TEST_VALUE_LIMIT = 50;
+};
+
+CV_SVMSGDTrainTest::CV_SVMSGDTrainTest(Mat weights, float shift)
+{
+    int datasize = 100000;
+    int varCount = weights.cols;
+    cv::Mat samples = cv::Mat::zeros( datasize, varCount, CV_32FC1 );
+    cv::Mat responses = cv::Mat::zeros( datasize, 1, CV_32FC1 );
+    cv::RNG rng(0);
+
+    float lowerLimit = -TEST_VALUE_LIMIT;
+    float upperLimit = TEST_VALUE_LIMIT;
+
+
+    rng.fill(samples, RNG::UNIFORM, lowerLimit, upperLimit);
+    for (int sampleIndex = 0; sampleIndex < datasize; sampleIndex++)
+    {
+        responses.at<float>( sampleIndex ) = decisionFunction(samples.row(sampleIndex), weights, shift) > 0 ? 1 : -1;
+    }
+
+    data = TrainData::create( samples, cv::ml::ROW_SAMPLE, responses );
+
+    int testSamplesCount = 100000;
+
+    testSamples.create(testSamplesCount, varCount, CV_32FC1);
+    rng.fill(testSamples, RNG::UNIFORM, lowerLimit, upperLimit);
+    testResponses.create(testSamplesCount, 1, CV_32FC1);
+
+    for (int i = 0 ; i < testSamplesCount; i++)
+    {
+        testResponses.at<float>(i) = decisionFunction(testSamples.row(i), weights, shift) > 0 ? 1 : -1;
+    }
+}
+
+void CV_SVMSGDTrainTest::run( int /*start_from*/ )
+{
+    cv::Ptr<SVMSGD> svmsgd = SVMSGD::create();
+
+    svmsgd->setOptimalParameters(SVMSGD::ASGD);
+
+    svmsgd->train( data );
+
+    Mat responses;
+
+    svmsgd->predict(testSamples, responses);
+
+    int errCount = 0;
+    int testSamplesCount = testSamples.rows;
+
+    for (int i = 0; i < testSamplesCount; i++)
+    {
+        if (responses.at<float>(i) * testResponses.at<float>(i) < 0 )
+            errCount++;
+    }
+
+    float err = (float)errCount / testSamplesCount;
+    std::cout << "err " << err << std::endl;
+
+    if ( err > 0.01 )
+    {
+        ts->set_failed_test_info( cvtest::TS::FAIL_BAD_ACCURACY );
+    }
+}
+
+float CV_SVMSGDTrainTest::decisionFunction(Mat sample, Mat weights, float shift)
+{
+    return sample.dot(weights) + shift;
+}
+
+TEST(ML_SVMSGD, train0)
+{
+    int varCount = 2;
+
+    Mat weights;
+    weights.create(1, varCount, CV_32FC1);
+    weights.at<float>(0) = 1;
+    weights.at<float>(1) = 0;
+
+    float shift = 5;
+
+    CV_SVMSGDTrainTest test(weights, shift);
+    test.safe_run();
+}
+
+TEST(ML_SVMSGD, train1)
+{
+    int varCount = 5;
+
+    Mat weights;
+    weights.create(1, varCount, CV_32FC1);
+
+    float lowerLimit = -1;
+    float upperLimit = 1;
+    cv::RNG rng(0);
+    rng.fill(weights, RNG::UNIFORM, lowerLimit, upperLimit);
+
+    float shift = rng.uniform(-5.f, 5.f);
+
+    CV_SVMSGDTrainTest test(weights, shift);
+    test.safe_run();
+}
+
+TEST(ML_SVMSGD, train2)
+{
+    int varCount = 100;
+
+    Mat weights;
+    weights.create(1, varCount, CV_32FC1);
+
+    float lowerLimit = -1;
+    float upperLimit = 1;
+    cv::RNG rng(0);
+    rng.fill(weights, RNG::UNIFORM, lowerLimit, upperLimit);
+
+    float shift = rng.uniform(-1000.f, 1000.f);
+
+    CV_SVMSGDTrainTest test(weights, shift);
+    test.safe_run();
+}