/// <summary> /// Postprocess the specified frame, outs and net. /// </summary> /// <param name="frame">Frame.</param> /// <param name="outs">Outs.</param> /// <param name="net">Net.</param> private void postprocess(Mat frame, List <Mat> outs, Net net) { string outLayerType = outBlobTypes[0]; List <int> classIdsList = new List <int>(); List <float> confidencesList = new List <float>(); List <OpenCVForUnity.CoreModule.Rect> boxesList = new List <OpenCVForUnity.CoreModule.Rect>(); if (net.getLayer(new DictValue(0)).outputNameToIndex("im_info") != -1) { // Faster-RCNN or R-FCN // Network produces output blob with a shape 1x1xNx7 where N is a number of // detections and an every detection is a vector of values // [batchId, classId, confidence, left, top, right, bottom] if (outs.Count == 1) { outs[0] = outs[0].reshape(1, (int)outs[0].total() / 7); //Debug.Log ("outs[i].ToString() " + outs [0].ToString ()); float[] data = new float[7]; for (int i = 0; i < outs[0].rows(); i++) { outs[0].get(i, 0, data); float confidence = data[2]; if (confidence > confThreshold) { int class_id = (int)(data[1]); int left = (int)(data[3] * frame.cols()); int top = (int)(data[4] * frame.rows()); int right = (int)(data[5] * frame.cols()); int bottom = (int)(data[6] * frame.rows()); int width = right - left + 1; int height = bottom - top + 1; classIdsList.Add((int)(class_id) - 0); confidencesList.Add((float)confidence); boxesList.Add(new OpenCVForUnity.CoreModule.Rect(left, top, width, height)); } } } } else if (outLayerType == "DetectionOutput") { // Network produces output blob with a shape 1x1xNx7 where N is a number of // detections and an every detection is a vector of values // [batchId, classId, confidence, left, top, right, bottom] if (outs.Count == 1) { outs[0] = outs[0].reshape(1, (int)outs[0].total() / 7); //Debug.Log ("outs[i].ToString() " + outs [0].ToString ()); float[] data = new float[7]; for (int i = 0; i < outs[0].rows(); i++) { outs[0].get(i, 0, data); float confidence = data[2]; if (confidence > confThreshold) { int class_id = (int)(data[1]); int left = (int)(data[3] * frame.cols()); int top = (int)(data[4] * frame.rows()); int right = (int)(data[5] * frame.cols()); int bottom = (int)(data[6] * frame.rows()); int width = right - left + 1; int height = bottom - top + 1; classIdsList.Add((int)(class_id) - 0); confidencesList.Add((float)confidence); boxesList.Add(new OpenCVForUnity.CoreModule.Rect(left, top, width, height)); } } } } else if (outLayerType == "Region") { for (int i = 0; i < outs.Count; ++i) { // Network produces output blob with a shape NxC where N is a number of // detected objects and C is a number of classes + 4 where the first 4 // numbers are [center_x, center_y, width, height] //Debug.Log ("outs[i].ToString() "+outs[i].ToString()); float[] positionData = new float[5]; float[] confidenceData = new float[outs[i].cols() - 5]; for (int p = 0; p < outs[i].rows(); p++) { outs[i].get(p, 0, positionData); outs[i].get(p, 5, confidenceData); int maxIdx = confidenceData.Select((val, idx) => new { V = val, I = idx }).Aggregate((max, working) => (max.V > working.V) ? max : working).I; float confidence = confidenceData[maxIdx]; if (confidence > confThreshold) { int centerX = (int)(positionData[0] * frame.cols()); int centerY = (int)(positionData[1] * frame.rows()); int width = (int)(positionData[2] * frame.cols()); int height = (int)(positionData[3] * frame.rows()); int left = centerX - width / 2; int top = centerY - height / 2; classIdsList.Add(maxIdx); confidencesList.Add((float)confidence); boxesList.Add(new OpenCVForUnity.CoreModule.Rect(left, top, width, height)); } } } } else { Debug.Log("Unknown output layer type: " + outLayerType); } MatOfRect boxes = new MatOfRect(); boxes.fromList(boxesList); MatOfFloat confidences = new MatOfFloat(); confidences.fromList(confidencesList); MatOfInt indices = new MatOfInt(); Dnn.NMSBoxes(boxes, confidences, confThreshold, nmsThreshold, indices); //Debug.Log ("indices.dump () "+indices.dump ()); //Debug.Log ("indices.ToString () "+indices.ToString()); for (int i = 0; i < indices.total(); ++i) { int idx = (int)indices.get(i, 0)[0]; OpenCVForUnity.CoreModule.Rect box = boxesList[idx]; drawPred(classIdsList[idx], confidencesList[idx], box.x, box.y, box.x + box.width, box.y + box.height, frame); } indices.Dispose(); boxes.Dispose(); confidences.Dispose(); }
/// <summary> /// Process /// </summary> /// <returns></returns> private async void Process() { float DOWNSCALE_RATIO = 1.0f; while (true) { // Check TaskCancel if (tokenSource.Token.IsCancellationRequested) { break; } rgbaMat = webCamTextureToMatHelper.GetMat(); // Debug.Log ("rgbaMat.ToString() " + rgbaMat.ToString ()); Mat downScaleRgbaMat = null; DOWNSCALE_RATIO = 1.0f; if (enableDownScale) { downScaleRgbaMat = imageOptimizationHelper.GetDownScaleMat(rgbaMat); DOWNSCALE_RATIO = imageOptimizationHelper.downscaleRatio; } else { downScaleRgbaMat = rgbaMat; DOWNSCALE_RATIO = 1.0f; } Imgproc.cvtColor(downScaleRgbaMat, bgrMat, Imgproc.COLOR_RGBA2BGR); await Task.Run(() => { // detect faces on the downscale image if (!enableSkipFrame || !imageOptimizationHelper.IsCurrentFrameSkipped()) { if (net == null) { Imgproc.putText(rgbaMat, "model file is not loaded.", new Point(5, rgbaMat.rows() - 30), Imgproc.FONT_HERSHEY_SIMPLEX, 0.7, new Scalar(255, 255, 255, 255), 2, Imgproc.LINE_AA, false); Imgproc.putText(rgbaMat, "Please read console message.", new Point(5, rgbaMat.rows() - 10), Imgproc.FONT_HERSHEY_SIMPLEX, 0.7, new Scalar(255, 255, 255, 255), 2, Imgproc.LINE_AA, false); } else { // Create a 4D blob from a frame. Size inpSize = new Size(inpWidth > 0 ? inpWidth : bgrMat.cols(), inpHeight > 0 ? inpHeight : bgrMat.rows()); Mat blob = Dnn.blobFromImage(bgrMat, scale, inpSize, mean, swapRB, false); // Run a model. net.setInput(blob); if (net.getLayer(new DictValue(0)).outputNameToIndex("im_info") != -1) { // Faster-RCNN or R-FCN Imgproc.resize(bgrMat, bgrMat, inpSize); Mat imInfo = new Mat(1, 3, CvType.CV_32FC1); imInfo.put(0, 0, new float[] { (float)inpSize.height, (float)inpSize.width, 1.6f }); net.setInput(imInfo, "im_info"); } TickMeter tm = new TickMeter(); tm.start(); List <Mat> outs = new List <Mat>(); net.forward(outs, outBlobNames); tm.stop(); // Debug.Log ("Inference time, ms: " + tm.getTimeMilli ()); postprocess(bgrMat, outs, net); for (int i = 0; i < outs.Count; i++) { outs[i].Dispose(); } blob.Dispose(); if (enableDownScale) { for (int i = 0; i < _boxesList.Count; ++i) { var rect = _boxesList[i]; _boxesList[i] = new OpenCVForUnity.CoreModule.Rect( (int)(rect.x * DOWNSCALE_RATIO), (int)(rect.y * DOWNSCALE_RATIO), (int)(rect.width * DOWNSCALE_RATIO), (int)(rect.height * DOWNSCALE_RATIO)); } } } //Imgproc.rectangle(rgbaMat, new Point(0, 0), new Point(rgbaMat.width(), rgbaMat.height()), new Scalar(0, 0, 0, 0), -1); MatOfRect boxes = new MatOfRect(); boxes.fromList(_boxesList); MatOfFloat confidences = new MatOfFloat(); confidences.fromList(_confidencesList); MatOfInt indices = new MatOfInt(); Dnn.NMSBoxes(boxes, confidences, confThreshold, nmsThreshold, indices); // Debug.Log ("indices.dump () "+indices.dump ()); // Debug.Log ("indices.ToString () "+indices.ToString()); for (int i = 0; i < indices.total(); ++i) { int idx = (int)indices.get(i, 0)[0]; OpenCVForUnity.CoreModule.Rect box = _boxesList[idx]; drawPred(_classIdsList[idx], _confidencesList[idx], box.x, box.y, box.x + box.width, box.y + box.height, rgbaMat); } indices.Dispose(); boxes.Dispose(); confidences.Dispose(); } }); Utils.fastMatToTexture2D(rgbaMat, texture); Thread.Sleep(10); } }
/// <summary> /// Postprocess the specified frame, outs and net. /// </summary> /// <param name="frame">Frame.</param> /// <param name="outs">Outs.</param> /// <param name="net">Net.</param> private void postprocess(Mat frame, List <Mat> outs, Net net) { string outLayerType = outBlobTypes[0]; List <int> classIdsList = new List <int>(); List <float> confidencesList = new List <float>(); List <OpenCVForUnity.CoreModule.Rect> boxesList = new List <OpenCVForUnity.CoreModule.Rect>(); if (outLayerType == "Region") { for (int i = 0; i < outs.Count; ++i) { // Network produces output blob with a shape NxC where N is a number of // detected objects and C is a number of classes + 4 where the first 4 // numbers are [center_x, center_y, width, height] //Debug.Log("outs[i].ToString() " + outs[i].ToString()); float[] positionData = new float[5]; float[] confidenceData = new float[outs[i].cols() - 5]; for (int p = 0; p < outs[i].rows(); p++) { outs[i].get(p, 0, positionData); outs[i].get(p, 5, confidenceData); int maxIdx = confidenceData.Select((val, idx) => new { V = val, I = idx }).Aggregate((max, working) => (max.V > working.V) ? max : working).I; float confidence = confidenceData[maxIdx]; if (confidence > confThreshold) { int centerX = (int)(positionData[0] * frame.cols()); int centerY = (int)(positionData[1] * frame.rows()); int width = (int)(positionData[2] * frame.cols()); int height = (int)(positionData[3] * frame.rows()); int left = centerX - width / 2; int top = centerY - height / 2; classIdsList.Add(maxIdx); confidencesList.Add((float)confidence); boxesList.Add(new OpenCVForUnity.CoreModule.Rect(left, top, width, height)); } } } } else { Debug.Log("Unknown output layer type: " + outLayerType); } MatOfRect boxes = new MatOfRect(); boxes.fromList(boxesList); MatOfFloat confidences = new MatOfFloat(); confidences.fromList(confidencesList); MatOfInt indices = new MatOfInt(); Dnn.NMSBoxes(boxes, confidences, confThreshold, nmsThreshold, indices); //Check the Language selected switch (menuVariables.GetLanguage()) { case "EN": vocOffset = 0; break; case "ES": vocOffset = 80; break; case "FR": vocOffset = 160; break; case "DE": vocOffset = 240; break; case "IT": vocOffset = 320; break; default: vocOffset = 0; break; } //Draw the bouding box only if its in the center of the image (On Cursor) for (int i = 0; i < indices.total(); ++i) { int idx = (int)indices.get(i, 0)[0]; OpenCVForUnity.CoreModule.Rect box = boxesList[idx]; if (isOnCursor(box, cursorObject.GetComponent <Cursor>())) { drawPred(vocOffset + classIdsList[idx], confidencesList[idx], box.x, box.y, box.x + box.width, box.y + box.height, frame); wordDisplay.text = classNames[classIdsList[idx]]; //Update the text summarizing the object encountered if (!vocIDList.Contains(classIdsList[idx])) { //Update the vocabulary learned vocIDList.Add(classIdsList[idx]); EnglishText.text += "\n" + classNames[classIdsList[idx]] + "\n"; SpanishText.text += "\n" + classNames[80 + classIdsList[idx]] + "\n"; FrenchText.text += "\n" + classNames[160 + classIdsList[idx]] + "\n"; GermanText.text += "\n" + classNames[240 + classIdsList[idx]] + "\n"; ItalianText.text += "\n" + classNames[320 + classIdsList[idx]] + "\n"; } } } indices.Dispose(); boxes.Dispose(); confidences.Dispose(); }
/// <summary> /// Scanning the specified frame, outs and net. /// </summary> /// <param name="frame">Frame.</param> /// <param name="outs">Outs.</param> /// <param name="net">Net.</param> private void postscan(Mat frame, List <Mat> outs, Net net) { string outLayerType = outBlobTypes[0]; List <int> classIdsList = new List <int>(); List <float> confidencesList = new List <float>(); List <OpenCVForUnity.CoreModule.Rect> boxesList = new List <OpenCVForUnity.CoreModule.Rect>(); if (outLayerType == "Region") { for (int i = 0; i < outs.Count; ++i) { // Network produces output blob with a shape NxC where N is a number of // detected objects and C is a number of classes + 4 where the first 4 // numbers are [center_x, center_y, width, height] //Debug.Log("outs[i].ToString() " + outs[i].ToString()); float[] positionData = new float[5]; float[] confidenceData = new float[outs[i].cols() - 5]; for (int p = 0; p < outs[i].rows(); p++) { outs[i].get(p, 0, positionData); outs[i].get(p, 5, confidenceData); int maxIdx = confidenceData.Select((val, idx) => new { V = val, I = idx }).Aggregate((max, working) => (max.V > working.V) ? max : working).I; float confidence = confidenceData[maxIdx]; if (confidence > confThreshold) { int centerX = (int)(positionData[0] * frame.cols()); int centerY = (int)(positionData[1] * frame.rows()); int width = (int)(positionData[2] * frame.cols()); int height = (int)(positionData[3] * frame.rows()); int left = centerX - width / 2; int top = centerY - height / 2; classIdsList.Add(maxIdx); confidencesList.Add((float)confidence); boxesList.Add(new OpenCVForUnity.CoreModule.Rect(left, top, width, height)); } } } } else { Debug.Log("Unknown output layer type: " + outLayerType); } MatOfRect boxes = new MatOfRect(); boxes.fromList(boxesList); MatOfFloat confidences = new MatOfFloat(); confidences.fromList(confidencesList); MatOfInt indices = new MatOfInt(); Dnn.NMSBoxes(boxes, confidences, confThreshold, nmsThreshold, indices); //for-loop for the mini game - if a new class appears, add it to the for (int i = 0; i < indices.total(); ++i) { int idx = (int)indices.get(i, 0)[0]; if (!minigameList.Contains(classIdsList[idx])) { Debug.Log(classNames[classIdsList[idx]]); minigameList.Add(classIdsList[idx]); if (minigameList.Count() > 1) { wordDisplay.text = minigameList.Count().ToString() + " words"; } else { wordDisplay.text = minigameList.Count().ToString() + " word"; } } } indices.Dispose(); boxes.Dispose(); confidences.Dispose(); }