improve code quality

- eliminate rand() calls
- non initialized members/ variables
- unused return values
- missing/useless NULL checks

modules/dnn/misc/python/pyopencv_dnn.hpp

@@ -142,7 +142,7 @@ public:
         PyGILState_Release(gstate);
         if (!res)
             CV_Error(Error::StsNotImplemented, "Failed to call \"getMemoryShapes\" method");
-        pyopencv_to_generic_vec(res, outputs, ArgInfo("", 0));
+        CV_Assert(pyopencv_to_generic_vec(res, outputs, ArgInfo("", 0)));
         return false;
     }
 
@@ -163,7 +163,7 @@ public:
             CV_Error(Error::StsNotImplemented, "Failed to call \"forward\" method");
 
         std::vector<Mat> pyOutputs;
-        pyopencv_to(res, pyOutputs, ArgInfo("", 0));
+        CV_Assert(pyopencv_to(res, pyOutputs, ArgInfo("", 0)));
 
         CV_Assert(pyOutputs.size() == outputs.size());
         for (size_t i = 0; i < outputs.size(); ++i)

modules/dnn/src/dnn.cpp

@@ -1530,10 +1530,12 @@ struct Net::Impl
                         LayerData *eltwiseData = nextData;
                         // go down from the second input and find the first non-skipped layer.
                         LayerData *downLayerData = &layers[eltwiseData->inputBlobsId[1].lid];
+                        CV_Assert(downLayerData);
                         while (downLayerData->skip)
                         {
                             downLayerData = &layers[downLayerData->inputBlobsId[0].lid];
                         }
+                        CV_Assert(downLayerData);
 
                         // second input layer is current layer.
                         if ( ld.id == downLayerData->id )
@@ -1548,9 +1550,7 @@ struct Net::Impl
                                     downLayerData = &layers[downLayerData->inputBlobsId[0].lid];
                             }
 
-                            Ptr<ConvolutionLayer> convLayer;
+                            Ptr<ConvolutionLayer> convLayer = downLayerData->layerInstance.dynamicCast<ConvolutionLayer>();
-                            if( downLayerData )
-                                convLayer = downLayerData->layerInstance.dynamicCast<ConvolutionLayer>();
 
                             //  first input layer is convolution layer
                             if( !convLayer.empty() && eltwiseData->consumers.size() == 1 )

modules/dnn/src/layers/recurrent_layers.cpp

@@ -119,9 +119,10 @@ public:
             if (blobs.size() > 3)
             {
                 CV_Assert(blobs.size() == 6);
+                const int N = Wh.cols;
                 for (int i = 3; i < 6; ++i)
                 {
-                    CV_Assert(blobs[i].rows == Wh.cols && blobs[i].cols == Wh.cols);
+                    CV_Assert(blobs[i].rows == N && blobs[i].cols == N);
                     CV_Assert(blobs[i].type() == bias.type());
                 }
             }

modules/dnn/src/ocl4dnn/src/math_functions.cpp

@@ -504,7 +504,7 @@ static bool ocl4dnnFastBufferGEMM(const CBLAS_TRANSPOSE TransA,
     oclk_gemm_float.set(arg_idx++, (float)alpha);
     oclk_gemm_float.set(arg_idx++, (float)beta);
 
-    bool ret;
+    bool ret = true;
     if (TransB == CblasNoTrans || TransA != CblasNoTrans) {
         int stride = 256;
         for (int start_index = 0; start_index < K; start_index += stride) {

modules/dnn/test/test_halide_layers.cpp

@@ -40,17 +40,18 @@ TEST(Padding_Halide, Accuracy)
 {
     static const int kNumRuns = 10;
     std::vector<int> paddings(8);
+    cv::RNG& rng = cv::theRNG();
     for (int t = 0; t < kNumRuns; ++t)
     {
         for (int i = 0; i < paddings.size(); ++i)
-            paddings[i] = rand() % 5;
+            paddings[i] = rng(5);
 
         LayerParams lp;
         lp.set("paddings", DictValue::arrayInt<int*>(&paddings[0], paddings.size()));
         lp.type = "Padding";
         lp.name = "testLayer";
 
-        Mat input({1 + rand() % 10, 1 + rand() % 10, 1 + rand() % 10, 1 + rand() % 10}, CV_32F);
+        Mat input({1 + rng(10), 1 + rng(10), 1 + rng(10), 1 + rng(10)}, CV_32F);
         test(lp, input);
     }
 }
@@ -633,7 +634,7 @@ TEST_P(Eltwise, Accuracy)
     eltwiseParam.set("operation", op);
     if (op == "sum" && weighted)
     {
-        RNG rng = cv::theRNG();
+        RNG& rng = cv::theRNG();
         std::vector<float> coeff(1 + numConv);
         for (int i = 0; i < coeff.size(); ++i)
         {

modules/dnn/test/test_tf_importer.cpp

@@ -376,7 +376,8 @@ TEST(Test_TensorFlow, memory_read)
 class ResizeBilinearLayer CV_FINAL : public Layer
 {
 public:
-    ResizeBilinearLayer(const LayerParams &params) : Layer(params)
+    ResizeBilinearLayer(const LayerParams &params) : Layer(params),
+        outWidth(0), outHeight(0), factorWidth(1), factorHeight(1)
     {
         CV_Assert(!params.get<bool>("align_corners", false));
         CV_Assert(!blobs.empty());

modules/imgproc/src/color.hpp

@@ -285,7 +285,8 @@ struct CvtHelper
 template< typename VScn, typename VDcn, typename VDepth, SizePolicy sizePolicy = NONE >
 struct OclHelper
 {
-    OclHelper( InputArray _src, OutputArray _dst, int dcn)
+    OclHelper( InputArray _src, OutputArray _dst, int dcn) :
+        nArgs(0)
     {
         src = _src.getUMat();
         Size sz = src.size(), dstSz;

modules/imgproc/test/test_intersection.cpp

@@ -357,20 +357,22 @@ void CV_RotatedRectangleIntersectionTest::test13()
 void CV_RotatedRectangleIntersectionTest::test14()
 {
     const int kNumTests = 100;
-    const int kWidth = 5;
+    const float kWidth = 5;
-    const int kHeight = 5;
+    const float kHeight = 5;
     RotatedRect rects[2];
     std::vector<Point2f> inter;
+    cv::RNG& rng = cv::theRNG();
     for (int i = 0; i < kNumTests; ++i)
     {
         for (int j = 0; j < 2; ++j)
         {
-            rects[j].center = Point2f((float)(rand() % kWidth), (float)(rand() % kHeight));
+            rects[j].center = Point2f(rng.uniform(0.0f, kWidth), rng.uniform(0.0f, kHeight));
-            rects[j].size = Size2f(rand() % kWidth + 1.0f, rand() % kHeight + 1.0f);
+            rects[j].size = Size2f(rng.uniform(1.0f, kWidth), rng.uniform(1.0f, kHeight));
-            rects[j].angle = (float)(rand() % 360);
+            rects[j].angle = rng.uniform(0.0f, 360.0f);
         }
-        rotatedRectangleIntersection(rects[0], rects[1], inter);
+        int res = rotatedRectangleIntersection(rects[0], rects[1], inter);
-        ASSERT_TRUE(inter.size() < 4 || isContourConvex(inter));
+        EXPECT_TRUE(res == INTERSECT_NONE || res == INTERSECT_PARTIAL || res == INTERSECT_FULL) << res;
+        ASSERT_TRUE(inter.size() < 4 || isContourConvex(inter)) << inter;
     }
 }