Fixed warnings.

modules/ml/include/opencv2/ml.hpp

@@ -1640,9 +1640,9 @@ public:
     CV_WRAP virtual void setOptimalParameters(int svmsgdType = SVMSGD::ASGD, int marginType = SVMSGD::SOFT_MARGIN) = 0;
 
     /** @brief %Algorithm type, one of SVMSGD::SvmsgdType. */
-    /** @see setAlgorithmType */
+    /** @see setSvmsgdType */
     CV_WRAP virtual int getSvmsgdType() const = 0;
-    /** @copybrief getAlgorithmType @see getAlgorithmType */
+    /** @copybrief getSvmsgdType @see getSvmsgdType */
     CV_WRAP virtual void setSvmsgdType(int svmsgdType) = 0;
 
     /** @brief %Margin type, one of SVMSGD::MarginType. */

modules/ml/src/svmsgd.cpp

@@ -172,7 +172,8 @@ void SVMSGDImpl::normalizeSamples(Mat &samples, Mat &average, float &multiplier)
     average = Mat(1, featuresCount, samples.type());
     for (int featureIndex = 0; featureIndex < featuresCount; featureIndex++)
     {
-        average.at<float>(featureIndex) = mean(samples.col(featureIndex))[0];
+        Scalar scalAverage = mean(samples.col(featureIndex))[0];
+        average.at<float>(featureIndex) = static_cast<float>(scalAverage[0]);
     }
 
     for (int sampleIndex = 0; sampleIndex < samplesCount; sampleIndex++)
@@ -182,7 +183,7 @@ void SVMSGDImpl::normalizeSamples(Mat &samples, Mat &average, float &multiplier)
 
     double normValue = norm(samples);
 
-    multiplier = sqrt(samples.total()) / normValue;
+    multiplier = static_cast<float>(sqrt(samples.total()) / normValue);
 
     samples *= multiplier;
 }
@@ -228,11 +229,11 @@ float SVMSGDImpl::calcShift(InputArray _samples, InputArray _responses) const
     for (int samplesIndex = 0; samplesIndex < trainSamplesCount; samplesIndex++)
     {
         Mat currentSample = trainSamples.row(samplesIndex);
-        float dotProduct = currentSample.dot(weights_);
+        float dotProduct = static_cast<float>(currentSample.dot(weights_));
 
         bool firstClass = isFirstClass(trainResponses.at<float>(samplesIndex));
-        int index = firstClass ? 0:1;
-        float signToMul = firstClass ? 1 : -1;
+        int index = firstClass ? 0 : 1;
+        float signToMul = firstClass ? 1.f : -1.f;
         float curDistance = dotProduct * signToMul;
 
         if (curDistance < distanceToClasses[index])
@@ -263,7 +264,7 @@ bool SVMSGDImpl::train(const Ptr<TrainData>& data, int)
     if ( areEmpty.first || areEmpty.second )
     {
         weights_ = Mat::zeros(1, featureCount, CV_32F);
-        shift_ = areEmpty.first ? -1 : 1;
+        shift_ = areEmpty.first ? -1.f : 1.f;
         return true;
     }
 
@@ -329,7 +330,7 @@ bool SVMSGDImpl::train(const Ptr<TrainData>& data, int)
 
     if (params.marginType == SOFT_MARGIN)
     {
-        shift_ = extendedWeights.at<float>(featureCount) - weights_.dot(average);
+        shift_ = extendedWeights.at<float>(featureCount) - static_cast<float>(weights_.dot(average));
     }
     else
     {
@@ -363,8 +364,8 @@ float SVMSGDImpl::predict( InputArray _samples, OutputArray _results, int ) cons
     for (int sampleIndex = 0; sampleIndex < nSamples; sampleIndex++)
     {
         Mat currentSample = samples.row(sampleIndex);
-        float criterion = currentSample.dot(weights_) + shift_;
-        results.at<float>(sampleIndex) = (criterion >= 0) ? 1 : -1;
+        float criterion = static_cast<float>(currentSample.dot(weights_)) + shift_;
+        results.at<float>(sampleIndex) = (criterion >= 0) ? 1.f : -1.f;
     }
 
     return result;
@@ -530,9 +531,9 @@ void SVMSGDImpl::setOptimalParameters(int svmsgdType, int marginType)
         params.svmsgdType = SGD;
         params.marginType = (marginType == SOFT_MARGIN) ? SOFT_MARGIN :
                             (marginType == HARD_MARGIN) ? HARD_MARGIN : ILLEGAL_MARGIN_TYPE;
-        params.lambda = 0.0001;
-        params.gamma0 = 0.05;
-        params.c = 1;
+        params.lambda = 0.0001f;
+        params.gamma0 = 0.05f;
+        params.c = 1.f;
         params.termCrit = TermCriteria(TermCriteria::COUNT + TermCriteria::EPS, 100000, 0.00001);
         break;
 
@@ -540,9 +541,9 @@ void SVMSGDImpl::setOptimalParameters(int svmsgdType, int marginType)
         params.svmsgdType = ASGD;
         params.marginType = (marginType == SOFT_MARGIN) ? SOFT_MARGIN :
                             (marginType == HARD_MARGIN) ? HARD_MARGIN : ILLEGAL_MARGIN_TYPE;
-        params.lambda = 0.00001;
-        params.gamma0 = 0.05;
-        params.c = 0.75;
+        params.lambda = 0.00001f;
+        params.gamma0 = 0.05f;
+        params.c = 0.75f;
         params.termCrit = TermCriteria(TermCriteria::COUNT + TermCriteria::EPS, 100000, 0.00001);
         break;
 

modules/ml/test/test_svmsgd.cpp

@@ -124,7 +124,7 @@ void CV_SVMSGDTrainTest::makeTrainData(Mat weights, float shift)
     cv::Mat responses = cv::Mat::zeros(datasize, 1, CV_32FC1);
     for (int sampleIndex = 0; sampleIndex < datasize; sampleIndex++)
     {
-        responses.at<float>(sampleIndex) = decisionFunction(samples.row(sampleIndex), weights, shift) > 0 ? 1 : -1;
+        responses.at<float>(sampleIndex) = decisionFunction(samples.row(sampleIndex), weights, shift) > 0 ? 1.f : -1.f;
     }
 
     data = TrainData::create(samples, cv::ml::ROW_SAMPLE, responses);
@@ -146,7 +146,7 @@ void CV_SVMSGDTrainTest::makeTestData(Mat weights, float shift)
 
     for (int i = 0 ; i < testSamplesCount; i++)
     {
-        testResponses.at<float>(i) = decisionFunction(testSamples.row(i), weights, shift) > 0 ? 1 : -1;
+        testResponses.at<float>(i) = decisionFunction(testSamples.row(i), weights, shift) > 0 ? 1.f : -1.f;
     }
 }
 
@@ -175,7 +175,7 @@ CV_SVMSGDTrainTest::CV_SVMSGDTrainTest(const Mat &weights, float shift, TrainDat
 
 float CV_SVMSGDTrainTest::decisionFunction(const Mat &sample, const Mat &weights, float shift)
 {
-    return sample.dot(weights) + shift;
+    return static_cast<float>(sample.dot(weights)) + shift;
 }
 
 void CV_SVMSGDTrainTest::run( int /*start_from*/ )
@@ -217,7 +217,7 @@ void makeWeightsAndShift(int featureCount, Mat &weights, float &shift)
     double upperLimit = 1;
 
     rng.fill(weights, RNG::UNIFORM, lowerLimit, upperLimit);
-    shift = rng.uniform(-featureCount, featureCount);
+    shift = static_cast<float>(rng.uniform(-featureCount, featureCount));
 }
 
 
@@ -319,7 +319,7 @@ TEST(ML_SVMSGD, twoPoints)
 
     float realShift = -500000.5;
 
-    float normRealWeights = norm(realWeights);
+    float normRealWeights = static_cast<float>(norm(realWeights));
     realWeights /= normRealWeights;
     realShift /= normRealWeights;
 
@@ -330,7 +330,7 @@ TEST(ML_SVMSGD, twoPoints)
     Mat foundWeights = svmsgd->getWeights();
     float foundShift = svmsgd->getShift();
 
-    float normFoundWeights = norm(foundWeights);
+    float normFoundWeights = static_cast<float>(norm(foundWeights));
     foundWeights /= normFoundWeights;
     foundShift /= normFoundWeights;
     CV_Assert((norm(foundWeights - realWeights) < 0.001) && (abs((foundShift - realShift) / realShift) < 0.05));

samples/cpp/train_svmsgd.cpp

@@ -97,7 +97,7 @@ bool findCrossPointWithBorders(const Mat &weights, float shift, const std::pair<
     if (xMin == xMax && weights.at<float>(1) != 0)
     {
         x = xMin;
-        y = std::floor( - (weights.at<float>(0) * x + shift) / weights.at<float>(1));
+        y = static_cast<int>(std::floor( - (weights.at<float>(0) * x + shift) / weights.at<float>(1)));
         if (y >= yMin && y <= yMax)
         {
             crossPoint.x = x;
@@ -108,7 +108,7 @@ bool findCrossPointWithBorders(const Mat &weights, float shift, const std::pair<
     else if (yMin == yMax && weights.at<float>(0) != 0)
     {
         y = yMin;
-        x = std::floor( - (weights.at<float>(1) * y + shift) / weights.at<float>(0));
+        x = static_cast<int>(std::floor( - (weights.at<float>(1) * y + shift) / weights.at<float>(0)));
         if (x >= xMin && x <= xMax)
         {
             crossPoint.x = x;
@@ -149,8 +149,8 @@ void redraw(Data data, const Point points[2])
     Scalar color;
     for (int i = 0; i < data.samples.rows; i++)
     {
-        center.x = data.samples.at<float>(i,0);
-        center.y = data.samples.at<float>(i,1);
+        center.x = static_cast<int>(data.samples.at<float>(i,0));
+        center.y = static_cast<int>(data.samples.at<float>(i,1));
         color = (data.responses.at<float>(i) > 0) ? Scalar(128,128,0) : Scalar(0,128,128);
         circle(data.img, center, radius, color, 5);
     }
@@ -163,8 +163,8 @@ void addPointRetrainAndRedraw(Data &data, int x, int y, int response)
 {
     Mat currentSample(1, 2, CV_32F);
 
-    currentSample.at<float>(0,0) = x;
-    currentSample.at<float>(0,1) = y;
+    currentSample.at<float>(0,0) = (float)x;
+    currentSample.at<float>(0,1) = (float)y;
     data.samples.push_back(currentSample);
     data.responses.push_back(response);