Merge remote-tracking branch 'upstream/3.4' into merge-3.4

cmake/OpenCVDetectInferenceEngine.cmake

@@ -138,8 +138,8 @@ if(INF_ENGINE_TARGET)
     math(EXPR INF_ENGINE_RELEASE "${InferenceEngine_VERSION_MAJOR} * 1000000 + ${InferenceEngine_VERSION_MINOR} * 10000 + ${InferenceEngine_VERSION_PATCH} * 100")
   endif()
   if(NOT INF_ENGINE_RELEASE)
-    message(WARNING "InferenceEngine version has not been set, 2021.3 will be used by default. Set INF_ENGINE_RELEASE variable if you experience build errors.")
-    set(INF_ENGINE_RELEASE "2021030000")
+    message(WARNING "InferenceEngine version has not been set, 2021.4 will be used by default. Set INF_ENGINE_RELEASE variable if you experience build errors.")
+    set(INF_ENGINE_RELEASE "2021040000")
   endif()
   set(INF_ENGINE_RELEASE "${INF_ENGINE_RELEASE}" CACHE STRING "Force IE version, should be in form YYYYAABBCC (e.g. 2020.1.0.2 -> 2020010002)")
   set_target_properties(${INF_ENGINE_TARGET} PROPERTIES

modules/core/include/opencv2/core/utils/filesystem.private.hpp

@@ -16,8 +16,8 @@
 #      define OPENCV_HAVE_FILESYSTEM_SUPPORT 1
 #  elif defined(__APPLE__)
 #    include <TargetConditionals.h>
-#    if (defined(TARGET_OS_OSX) && TARGET_OS_OSX) || (!defined(TARGET_OS_OSX) && !TARGET_OS_IPHONE)
-#      define OPENCV_HAVE_FILESYSTEM_SUPPORT 1 // OSX only
+#    if (defined(TARGET_OS_OSX) && TARGET_OS_OSX) || (defined(TARGET_OS_IOS) && TARGET_OS_IOS)
+#      define OPENCV_HAVE_FILESYSTEM_SUPPORT 1 // OSX, iOS only
 #    endif
 #  else
      /* unknown */

modules/dnn/src/ie_ngraph.cpp

@@ -657,7 +657,11 @@ void InfEngineNgraphNet::initPlugin(InferenceEngine::CNNNetwork& net)
                 try
                 {
                     InferenceEngine::IExtensionPtr extension =
+#if INF_ENGINE_VER_MAJOR_GE(INF_ENGINE_RELEASE_2021_4)
+                        std::make_shared<InferenceEngine::Extension>(libName);
+#else
                         InferenceEngine::make_so_pointer<InferenceEngine::IExtension>(libName);
+#endif
 
                     ie.AddExtension(extension, "CPU");
                     CV_LOG_INFO(NULL, "DNN-IE: Loaded extension plugin: " << libName);
@@ -1005,35 +1009,54 @@ void InfEngineNgraphNet::forward(const std::vector<Ptr<BackendWrapper> >& outBlo
         reqWrapper->req.SetInput(inpBlobs);
         reqWrapper->req.SetOutput(outBlobs);
 
+#if INF_ENGINE_VER_MAJOR_GE(INF_ENGINE_RELEASE_2021_4)
+        InferenceEngine::InferRequest infRequest = reqWrapper->req;
+        NgraphReqWrapper* wrapperPtr = reqWrapper.get();
+        CV_Assert(wrapperPtr && "Internal error");
+#else
         InferenceEngine::IInferRequest::Ptr infRequestPtr = reqWrapper->req;
-        infRequestPtr->SetUserData(reqWrapper.get(), 0);
+        CV_Assert(infRequestPtr);
+        InferenceEngine::IInferRequest& infRequest = *infRequestPtr.get();
+        infRequest.SetUserData(reqWrapper.get(), 0);
+#endif
 
-        infRequestPtr->SetCompletionCallback(
-            [](InferenceEngine::IInferRequest::Ptr request, InferenceEngine::StatusCode status)
+#if INF_ENGINE_VER_MAJOR_GE(INF_ENGINE_RELEASE_2021_4)
+        // do NOT capture 'reqWrapper' (smart ptr) in the lambda callback
+        infRequest.SetCompletionCallback<std::function<void(InferenceEngine::InferRequest, InferenceEngine::StatusCode)>>(
+            [wrapperPtr](InferenceEngine::InferRequest /*request*/, InferenceEngine::StatusCode status)
+#else
+        infRequest.SetCompletionCallback(
+            [](InferenceEngine::IInferRequest::Ptr requestPtr, InferenceEngine::StatusCode status)
+#endif
             {
                 CV_LOG_DEBUG(NULL, "DNN(nGraph): completionCallback(" << (int)status << ")");
+#if !INF_ENGINE_VER_MAJOR_GE(INF_ENGINE_RELEASE_2021_4)
+                CV_Assert(requestPtr);
+                InferenceEngine::IInferRequest& request = *requestPtr.get();
 
-                NgraphReqWrapper* wrapper;
-                request->GetUserData((void**)&wrapper, 0);
-                CV_Assert(wrapper && "Internal error");
+                NgraphReqWrapper* wrapperPtr;
+                request.GetUserData((void**)&wrapperPtr, 0);
+                CV_Assert(wrapperPtr && "Internal error");
+#endif
+                NgraphReqWrapper& wrapper = *wrapperPtr;
 
                 size_t processedOutputs = 0;
                 try
                 {
-                    for (; processedOutputs < wrapper->outProms.size(); ++processedOutputs)
+                    for (; processedOutputs < wrapper.outProms.size(); ++processedOutputs)
                     {
-                        const std::string& name = wrapper->outsNames[processedOutputs];
-                        Mat m = ngraphBlobToMat(wrapper->req.GetBlob(name));
+                        const std::string& name = wrapper.outsNames[processedOutputs];
+                        Mat m = ngraphBlobToMat(wrapper.req.GetBlob(name));
 
                         try
                         {
                             CV_Assert(status == InferenceEngine::StatusCode::OK);
-                            wrapper->outProms[processedOutputs].setValue(m.clone());
+                            wrapper.outProms[processedOutputs].setValue(m.clone());
                         }
                         catch (...)
                         {
                             try {
-                                wrapper->outProms[processedOutputs].setException(std::current_exception());
+                                wrapper.outProms[processedOutputs].setException(std::current_exception());
                             } catch(...) {
                                 CV_LOG_ERROR(NULL, "DNN: Exception occurred during async inference exception propagation");
                             }
@@ -1043,16 +1066,16 @@ void InfEngineNgraphNet::forward(const std::vector<Ptr<BackendWrapper> >& outBlo
                 catch (...)
                 {
                     std::exception_ptr e = std::current_exception();
-                    for (; processedOutputs < wrapper->outProms.size(); ++processedOutputs)
+                    for (; processedOutputs < wrapper.outProms.size(); ++processedOutputs)
                     {
                         try {
-                            wrapper->outProms[processedOutputs].setException(e);
+                            wrapper.outProms[processedOutputs].setException(e);
                         } catch(...) {
                             CV_LOG_ERROR(NULL, "DNN: Exception occurred during async inference exception propagation");
                         }
                     }
                 }
-                wrapper->isReady = true;
+                wrapper.isReady = true;
             }
         );
     }

modules/dnn/src/layers/batch_norm_layer.cpp

@@ -35,6 +35,7 @@ namespace dnn
 class BatchNormLayerImpl CV_FINAL : public BatchNormLayer
 {
 public:
+    Mat origin_weights, origin_bias;
     Mat weights_, bias_;
     UMat umat_weight, umat_bias;
     mutable int dims;
@@ -88,11 +89,11 @@ public:
         const float* weightsData = hasWeights ? blobs[weightsBlobIndex].ptr<float>() : 0;
         const float* biasData = hasBias ? blobs[biasBlobIndex].ptr<float>() : 0;
 
-        weights_.create(1, (int)n, CV_32F);
-        bias_.create(1, (int)n, CV_32F);
+        origin_weights.create(1, (int)n, CV_32F);
+        origin_bias.create(1, (int)n, CV_32F);
 
-        float* dstWeightsData = weights_.ptr<float>();
-        float* dstBiasData = bias_.ptr<float>();
+        float* dstWeightsData = origin_weights.ptr<float>();
+        float* dstBiasData = origin_bias.ptr<float>();
 
         for (size_t i = 0; i < n; ++i)
         {
@@ -100,15 +101,12 @@ public:
             dstWeightsData[i] = w;
             dstBiasData[i] = (hasBias ? biasData[i] : 0.0f) - w * meanData[i] * varMeanScale;
         }
-        // We will use blobs to store origin weights and bias to restore them in case of reinitialization.
-        weights_.copyTo(blobs[0].reshape(1, 1));
-        bias_.copyTo(blobs[1].reshape(1, 1));
     }
 
     virtual void finalize(InputArrayOfArrays, OutputArrayOfArrays) CV_OVERRIDE
     {
-        blobs[0].reshape(1, 1).copyTo(weights_);
-        blobs[1].reshape(1, 1).copyTo(bias_);
+        origin_weights.reshape(1, 1).copyTo(weights_);
+        origin_bias.reshape(1, 1).copyTo(bias_);
     }
 
     void getScaleShift(Mat& scale, Mat& shift) const CV_OVERRIDE

modules/dnn/src/onnx/onnx_importer.cpp

@@ -1954,6 +1954,23 @@ void ONNXImporter::handleNode(const opencv_onnx::NodeProto& node_proto_)
                 addConstant(layerParams.name, concatenated[0]);
                 return;
             }
+            else
+            {
+                for (int i = 0; i < node_proto.input_size(); ++i)
+                {
+                    if (constBlobs.find(node_proto.input(i)) != constBlobs.end())
+                    {
+                        LayerParams constParams;
+                        constParams.name = node_proto.input(i);
+                        constParams.type = "Const";
+                        constParams.blobs.push_back(getBlob(node_proto, i));
+
+                        opencv_onnx::NodeProto proto;
+                        proto.add_output(constParams.name);
+                        addLayer(constParams, proto);
+                    }
+                }
+            }
         }
         else if (layer_type == "Resize")
         {

modules/dnn/src/op_inf_engine.hpp

@@ -30,10 +30,11 @@
 #define INF_ENGINE_RELEASE_2021_1 2021010000
 #define INF_ENGINE_RELEASE_2021_2 2021020000
 #define INF_ENGINE_RELEASE_2021_3 2021030000
+#define INF_ENGINE_RELEASE_2021_4 2021040000
 
 #ifndef INF_ENGINE_RELEASE
-#warning("IE version have not been provided via command-line. Using 2021.3 by default")
-#define INF_ENGINE_RELEASE INF_ENGINE_RELEASE_2021_3
+#warning("IE version have not been provided via command-line. Using 2021.4 by default")
+#define INF_ENGINE_RELEASE INF_ENGINE_RELEASE_2021_4
 #endif
 
 #define INF_ENGINE_VER_MAJOR_GT(ver) (((INF_ENGINE_RELEASE) / 10000) > ((ver) / 10000))

modules/dnn/test/test_backends.cpp

@@ -204,7 +204,7 @@ TEST_P(DNNTestNetwork, MobileNet_SSD_Caffe)
     Mat inp = blobFromImage(sample, 1.0f / 127.5, Size(300, 300), Scalar(127.5, 127.5, 127.5), false);
     float scoreDiff = (target == DNN_TARGET_OPENCL_FP16 || target == DNN_TARGET_MYRIAD) ? 1.5e-2 : 0.0;
     float iouDiff = (target == DNN_TARGET_MYRIAD) ? 0.063  : 0.0;
-    float detectionConfThresh = (target == DNN_TARGET_MYRIAD) ? 0.252  : FLT_MIN;
+    float detectionConfThresh = (target == DNN_TARGET_MYRIAD) ? 0.262  : FLT_MIN;
          processNet("dnn/MobileNetSSD_deploy.caffemodel", "dnn/MobileNetSSD_deploy.prototxt",
                     inp, "detection_out", "", scoreDiff, iouDiff, detectionConfThresh);
     expectNoFallbacksFromIE(net);
@@ -359,8 +359,8 @@ TEST_P(DNNTestNetwork, OpenPose_pose_coco)
         applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD_X, CV_TEST_TAG_DNN_SKIP_IE_VERSION);
 #endif
 
-    const float l1 = (target == DNN_TARGET_MYRIAD) ? 0.0056 : 0.0;
-    const float lInf = (target == DNN_TARGET_MYRIAD) ? 0.072 : 0.0;
+    const float l1 = (target == DNN_TARGET_MYRIAD) ? 0.009 : 0.0;
+    const float lInf = (target == DNN_TARGET_MYRIAD) ? 0.09 : 0.0;
     processNet("dnn/openpose_pose_coco.caffemodel", "dnn/openpose_pose_coco.prototxt",
                Size(46, 46), "", "", l1, lInf);
     expectNoFallbacksFromIE(net);
@@ -380,8 +380,8 @@ TEST_P(DNNTestNetwork, OpenPose_pose_mpi)
 #endif
 
     // output range: [-0.001, 0.97]
-    const float l1 = (target == DNN_TARGET_MYRIAD) ? 0.012 : 0.0;
-    const float lInf = (target == DNN_TARGET_MYRIAD || target == DNN_TARGET_OPENCL_FP16) ? 0.16 : 0.0;
+    const float l1 = (target == DNN_TARGET_MYRIAD) ? 0.02 : 0.0;
+    const float lInf = (target == DNN_TARGET_MYRIAD || target == DNN_TARGET_OPENCL_FP16) ? 0.2 : 0.0;
     processNet("dnn/openpose_pose_mpi.caffemodel", "dnn/openpose_pose_mpi.prototxt",
                Size(46, 46), "", "", l1, lInf);
     expectNoFallbacksFromIE(net);

modules/dnn/test/test_ie_models.cpp

@@ -307,6 +307,15 @@ TEST_P(DNNTestOpenVINO, models)
     ASSERT_FALSE(backendId != DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 && backendId != DNN_BACKEND_INFERENCE_ENGINE_NGRAPH) <<
         "Inference Engine backend is required";
 
+#if INF_ENGINE_VER_MAJOR_EQ(2021040000)
+    if (targetId == DNN_TARGET_MYRIAD && (
+            modelName == "person-detection-retail-0013" ||  // ncDeviceOpen:1013 Failed to find booted device after boot
+            modelName == "age-gender-recognition-retail-0013"  // ncDeviceOpen:1013 Failed to find booted device after boot
+        )
+    )
+        applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD, CV_DNN_BACKEND_INFERENCE_ENGINE_NGRAPH, CV_TEST_TAG_DNN_SKIP_IE_VERSION);
+#endif
+
 #if INF_ENGINE_VER_MAJOR_GE(2020020000)
     if (targetId == DNN_TARGET_MYRIAD && backendId == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019)
     {

modules/dnn/test/test_onnx_importer.cpp

@@ -349,6 +349,7 @@ TEST_P(Test_ONNX_layers, Concatenation)
         if (target == DNN_TARGET_MYRIAD)      applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD, CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER);
     }
     testONNXModels("concatenation");
+    testONNXModels("concat_const_blobs");
 }
 
 TEST_P(Test_ONNX_layers, Eltwise3D)

modules/dnn/test/test_torch_importer.cpp

@@ -290,9 +290,14 @@ TEST_P(Test_Torch_layers, net_padding)
 
 TEST_P(Test_Torch_layers, net_non_spatial)
 {
-#if defined(INF_ENGINE_RELEASE) && INF_ENGINE_VER_MAJOR_EQ(2021030000)
+#if defined(INF_ENGINE_RELEASE) && ( \
+    INF_ENGINE_VER_MAJOR_EQ(2021030000) || \
+    INF_ENGINE_VER_MAJOR_EQ(2021040000) \
+)
     if (backend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH && target == DNN_TARGET_MYRIAD)
-        applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD, CV_TEST_TAG_DNN_SKIP_IE_NGRAPH);  // crash
+        // 2021.3: crash
+        // 2021.4: [ GENERAL_ERROR ]  AssertionFailed: !out.networkInputs.empty()
+        applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD, CV_TEST_TAG_DNN_SKIP_IE_NGRAPH);
     if (backend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH && target == DNN_TARGET_OPENCL)
         applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_OPENCL, CV_TEST_TAG_DNN_SKIP_IE_NGRAPH);  // exception
     if (backend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH && target == DNN_TARGET_OPENCL_FP16)

samples/cpp/tutorial_code/Histograms_Matching/MatchTemplate_Demo.cpp

@@ -89,7 +89,7 @@ void MatchingMethod( int, void* )
 
   //! [create_result_matrix]
   /// Create the result matrix
-  int result_cols =  img.cols - templ.cols + 1;
+  int result_cols = img.cols - templ.cols + 1;
   int result_rows = img.rows - templ.rows + 1;
 
   result.create( result_rows, result_cols, CV_32FC1 );

samples/cpp/tutorial_code/Histograms_Matching/calcBackProject_Demo1.cpp

@@ -72,18 +72,18 @@ void Hist_and_Backproj(int, void* )
     //! [initialize]
     int histSize = MAX( bins, 2 );
     float hue_range[] = { 0, 180 };
-    const float* ranges = { hue_range };
+    const float* ranges[] = { hue_range };
     //! [initialize]
 
     //! [Get the Histogram and normalize it]
     Mat hist;
-    calcHist( &hue, 1, 0, Mat(), hist, 1, &histSize, &ranges, true, false );
+    calcHist( &hue, 1, 0, Mat(), hist, 1, &histSize, ranges, true, false );
     normalize( hist, hist, 0, 255, NORM_MINMAX, -1, Mat() );
     //! [Get the Histogram and normalize it]
 
     //! [Get Backprojection]
     Mat backproj;
-    calcBackProject( &hue, 1, 0, hist, backproj, &ranges, 1, true );
+    calcBackProject( &hue, 1, 0, hist, backproj, ranges, 1, true );
     //! [Get Backprojection]
 
     //! [Draw the backproj]

samples/cpp/tutorial_code/Histograms_Matching/calcHist_Demo.cpp

@@ -37,7 +37,7 @@ int main(int argc, char** argv)
 
     //! [Set the ranges ( for B,G,R) )]
     float range[] = { 0, 256 }; //the upper boundary is exclusive
-    const float* histRange = { range };
+    const float* histRange[] = { range };
     //! [Set the ranges ( for B,G,R) )]
 
     //! [Set histogram param]
@@ -46,9 +46,9 @@ int main(int argc, char** argv)
 
     //! [Compute the histograms]
     Mat b_hist, g_hist, r_hist;
-    calcHist( &bgr_planes[0], 1, 0, Mat(), b_hist, 1, &histSize, &histRange, uniform, accumulate );
-    calcHist( &bgr_planes[1], 1, 0, Mat(), g_hist, 1, &histSize, &histRange, uniform, accumulate );
-    calcHist( &bgr_planes[2], 1, 0, Mat(), r_hist, 1, &histSize, &histRange, uniform, accumulate );
+    calcHist( &bgr_planes[0], 1, 0, Mat(), b_hist, 1, &histSize, histRange, uniform, accumulate );
+    calcHist( &bgr_planes[1], 1, 0, Mat(), g_hist, 1, &histSize, histRange, uniform, accumulate );
+    calcHist( &bgr_planes[2], 1, 0, Mat(), r_hist, 1, &histSize, histRange, uniform, accumulate );
     //! [Compute the histograms]
 
     //! [Draw the histograms for B, G and R]

samples/python/camera_calibration_show_extrinsics.py

@@ -188,7 +188,7 @@ def main():
 
     fig = plt.figure()
     ax = fig.gca(projection='3d')
-    ax.set_aspect("equal")
+    ax.set_aspect("auto")
 
     cam_width = args.cam_width
     cam_height = args.cam_height

samples/python/gaussian_mix.py

@@ -32,7 +32,7 @@ def draw_gaussain(img, mean, cov, color):
     w, u, _vt = cv.SVDecomp(cov)
     ang = np.arctan2(u[1, 0], u[0, 0])*(180/np.pi)
     s1, s2 = np.sqrt(w)*3.0
-    cv.ellipse(img, (x, y), (s1, s2), ang, 0, 360, color, 1, cv.LINE_AA)
+    cv.ellipse(img, (int(x), int(y)), (int(s1), int(s2)), ang, 0, 360, color, 1, cv.LINE_AA)
 
 
 def main():

samples/python/hist.py

@@ -48,7 +48,7 @@ def hist_lines(im):
     cv.normalize(hist_item,hist_item,0,255,cv.NORM_MINMAX)
     hist=np.int32(np.around(hist_item))
     for x,y in enumerate(hist):
-        cv.line(h,(x,0),(x,y),(255,255,255))
+        cv.line(h,(x,0),(x,y[0]),(255,255,255))
     y = np.flipud(h)
     return y
 

samples/python/lk_homography.py

@@ -77,8 +77,8 @@ class App:
 
                 for (x0, y0), (x1, y1), good in zip(self.p0[:,0], self.p1[:,0], status[:,0]):
                     if good:
-                        cv.line(vis, (x0, y0), (x1, y1), (0, 128, 0))
-                    cv.circle(vis, (x1, y1), 2, (red, green)[good], -1)
+                        cv.line(vis, (int(x0), int(y0)), (int(x1), int(y1)), (0, 128, 0))
+                    cv.circle(vis, (int(x1), int(y1)), 2, (red, green)[good], -1)
                 draw_str(vis, (20, 20), 'track count: %d' % len(self.p1))
                 if self.use_ransac:
                     draw_str(vis, (20, 40), 'RANSAC')
@@ -86,7 +86,7 @@ class App:
                 p = cv.goodFeaturesToTrack(frame_gray, **feature_params)
                 if p is not None:
                     for x, y in p[:,0]:
-                        cv.circle(vis, (x, y), 2, green, -1)
+                        cv.circle(vis, (int(x), int(y)), 2, green, -1)
                     draw_str(vis, (20, 20), 'feature count: %d' % len(p))
 
             cv.imshow('lk_homography', vis)

samples/python/lk_track.py

@@ -65,7 +65,7 @@ class App:
                     if len(tr) > self.track_len:
                         del tr[0]
                     new_tracks.append(tr)
-                    cv.circle(vis, (x, y), 2, (0, 255, 0), -1)
+                    cv.circle(vis, (int(x), int(y)), 2, (0, 255, 0), -1)
                 self.tracks = new_tracks
                 cv.polylines(vis, [np.int32(tr) for tr in self.tracks], False, (0, 255, 0))
                 draw_str(vis, (20, 20), 'track count: %d' % len(self.tracks))

samples/python/video_v4l2.py

@@ -30,7 +30,7 @@ def main():
     color = (0, 255, 0)
 
     cap = cv.VideoCapture(0)
-    cap.set(cv.CAP_PROP_AUTOFOCUS, False)  # Known bug: https://github.com/opencv/opencv/pull/5474
+    cap.set(cv.CAP_PROP_AUTOFOCUS, 0)  # Known bug: https://github.com/opencv/opencv/pull/5474
 
     cv.namedWindow("Video")
 
@@ -67,7 +67,7 @@ def main():
             break
         elif k == ord('g'):
             convert_rgb = not convert_rgb
-            cap.set(cv.CAP_PROP_CONVERT_RGB, convert_rgb)
+            cap.set(cv.CAP_PROP_CONVERT_RGB, 1 if convert_rgb else 0)
 
     print('Done')