OpenCV4實現Mask-RCNN

學習OpenCV4很好的一個英文博客和代碼示例：https://github.com/spmallick/learnopencv

本文使用的模型文件、數據等均可以在上面下載得到。

1、代碼配置

使用VS2015和OpenCV4.0.0實現，僅使用CPU(也可以使用GPU，根據電腦配置決定)。在Linux系統下編譯，同時使用GPU加速最好。

2、代碼實現

該代碼是根據上面網址的代碼做簡單的整理

mask_rcnn.h:

#pragma once

#include <fstream>
#include <sstream>
#include <iostream>
#include <string.h>

#include <opencv2/dnn.hpp>
#include <opencv2/imgproc.hpp>
#include <opencv2/highgui.hpp>

using namespace cv;
using namespace std;

class mask_rcnn
{
public:
	mask_rcnn(float confThreshold = 0.5, float maskThreshold = 0.3);

	/** @brief Draw the predicted bounding box, colorize and show the mask on the image
	* @param[in] image or image frame in video stream
	* @param[in] the index value corresponding to the category
	* @param[in] bounding box
	* @param[in] mask
	* @returns NULL.
	*/
	void drawBox(cv::Mat& frame, int classId, float conf, cv::Rect box, cv::Mat& objectMask);

	/** @brief Postprocess the neural network's output for each frame.For each frame, extract the bounding box and mask for each detected object
	* @param[in]  image or image frame in video stream
	* @param[out] Output size of masks is NxCxHxW where
    *             N - number of detected boxes
    *             C - number of classes (excluding background)
    *             HxW - segmentation shape
	* @returns NULL.
	*/
	void postprocess(cv::Mat& frame, const std::vector<cv::Mat>& outs);

	/** @brief detect object and mask in the image
	* @param[in] Load names of classes
	* @param[in] Load the colors
	* @param[in] Give the configuration 
	* @param[in] Give weight files for the model
	* @returns NULL.
	*/
	void detect_image(std::string classesFile, std::string colorsFile, String textGraph, String modelWeights, std::string imagepath, std::string& outputFile);

	/** @brief detect object and mask in the video
	* @param[in] Load names of classes
	* @param[in] Load the colors
	* @param[in] Give the configuration
	* @param[in] Give weight files for the model
	* @returns NULL.
	*/
	void detect_video(std::string classesFile, std::string colorsFile, String textGraph, String modelWeights, std::string videopath, std::string& outputFile);

private:
	// Initialize the parameters
	float mfConfThreshold;              // Confidence threshold
	float mfMaskThreshold;              // Mask threshold

	std::vector<std::string> mvClasses; // Recognizable object categories
	std::vector<cv::Scalar> mvColors;   // Recognition box color set
};

mask_rcnn.cpp:

#include "mask_rcnn.h"

mask_rcnn::mask_rcnn(float confThreshold, float maskThreshold)
{
	mfConfThreshold = confThreshold;
	mfMaskThreshold = maskThreshold;
}

void mask_rcnn::detect_image(std::string classesFile, std::string colorsFile, String textGraph, String modelWeights, std::string imagepath, std::string& outputFile)
{
	// Load names of classes
	std::string line;
	ifstream ifs(classesFile.c_str());
	while (getline(ifs, line)) mvClasses.push_back(line);

	// Load the colors
	ifstream colorFptr(colorsFile.c_str());
	while (getline(colorFptr, line))
	{
		char* pEnd;
		double r, g, b;
		r = strtod(line.c_str(), &pEnd);
		g = strtod(pEnd, NULL);
		b = strtod(pEnd, NULL);
		Scalar color = Scalar(r, g, b, 255.0);
		mvColors.push_back(Scalar(r, g, b, 255.0));
	}

	// Create a window
	static const string kWinName = "Deep learning object detection in OpenCV";
	namedWindow(kWinName, WINDOW_NORMAL);

	// Load the network
	dnn::Net net = dnn::readNetFromTensorflow(modelWeights, textGraph);
	net.setPreferableBackend(dnn::DNN_BACKEND_OPENCV);
	net.setPreferableTarget(dnn::DNN_TARGET_CPU);

	// Create a 4D blob from a frame.
	cv::Mat blob;
	cv::Mat frame = cv::imread(imagepath);

	// Create a 4D blob from a frame.
	dnn::blobFromImage(frame, blob, 1.0, Size(frame.cols, frame.rows), Scalar(), true, false);

	//Sets the input to the network
	net.setInput(blob);

	// Runs the forward pass to get output from the output layers
	std::vector<String> outNames(2);
	outNames[0] = "detection_out_final";
	outNames[1] = "detection_masks";
	vector<Mat> outs;
	net.forward(outs, outNames);

	// Extract the bounding box and mask for each of the detected objects
	postprocess(frame, outs);

	// Put efficiency information. The function getPerfProfile returns the overall time for inference(t) and the timings for each of the layers(in layersTimes)
	vector<double> layersTimes;
	double freq = getTickFrequency() / 1000;
	double t = net.getPerfProfile(layersTimes) / freq;
	string label = format("Mask-RCNN on 2.2 GHz Intel Core E5 CPU, Inference time for a frame : %0.0f ms", t);
	putText(frame, label, Point(0, 15), FONT_HERSHEY_SIMPLEX, 0.5, Scalar(0, 0, 0));

	// Write the frame with the detection boxes
	imshow(kWinName, frame);
	cv::imwrite(outputFile, frame);
}

void mask_rcnn::detect_video(std::string classesFile, std::string colorsFile, String textGraph, String modelWeights, std::string videopath, std::string& outputFile)
{
	// Load names of classes
	std::string line;
	ifstream ifs(classesFile.c_str());
	while (getline(ifs, line)) mvClasses.push_back(line);

	// Load the colors
	ifstream colorFptr(colorsFile.c_str());
	while (getline(colorFptr, line)) 
	{
		char* pEnd;
		double r, g, b;
		r = strtod(line.c_str(), &pEnd);
		g = strtod(pEnd, NULL);
		b = strtod(pEnd, NULL);
		Scalar color = Scalar(r, g, b, 255.0);
		mvColors.push_back(Scalar(r, g, b, 255.0));
	}

	// Load the network
	dnn::Net net = dnn::readNetFromTensorflow(modelWeights, textGraph);
	net.setPreferableBackend(dnn::DNN_BACKEND_OPENCV);
	net.setPreferableTarget(dnn::DNN_TARGET_CPU);

	// Open a video file or an image file or a camera stream.
	VideoCapture cap;
	VideoWriter video;

	Mat frame, blob;

	try {
		// Open the video file
		ifstream ifile(videopath);
		if (!ifile) throw("error");
		cap.open(videopath);
	}
	catch (...) {
		cout << "Could not open the input image/video stream" << endl;
		return;
	}

	// Get the video writer initialized to save the output video
	video.open(outputFile, VideoWriter::fourcc('M', 'J', 'P', 'G'), 28, Size(cap.get(CAP_PROP_FRAME_WIDTH), cap.get(CAP_PROP_FRAME_HEIGHT)));

	// Create a window
	static const string kWinName = "Deep learning object detection in OpenCV";
	namedWindow(kWinName, WINDOW_NORMAL);

	// Process frames.
	while (waitKey(1) < 0)
	{
		// get frame from the video
		cap >> frame;

		// Stop the program if reached end of video
		if (frame.empty()) {
			cout << "Done processing !!!" << endl;
			cout << "Output file is stored as " << outputFile << endl;
			waitKey(3000);
			break;
		}

		// Create a 4D blob from a frame.
		dnn::blobFromImage(frame, blob, 1.0, Size(frame.cols, frame.rows), Scalar(), true, false);

		//Sets the input to the network
		net.setInput(blob);

		// Runs the forward pass to get output from the output layers
		std::vector<String> outNames(2);
		outNames[0] = "detection_out_final";
		outNames[1] = "detection_masks";
		vector<Mat> outs;
		net.forward(outs, outNames);

		// Extract the bounding box and mask for each of the detected objects
		postprocess(frame, outs);

		// Put efficiency information. The function getPerfProfile returns the overall time for inference(t) and the timings for each of the layers(in layersTimes)
		vector<double> layersTimes;
		double freq = getTickFrequency() / 1000;
		double t = net.getPerfProfile(layersTimes) / freq;
		string label = format("Mask-RCNN on 2.2 GHz Intel Core E5 CPU, Inference time for a frame : %0.0f ms", t);
		putText(frame, label, Point(0, 15), FONT_HERSHEY_SIMPLEX, 0.5, Scalar(0, 0, 0));

		// Write the frame with the detection boxes
		Mat detectedFrame;
		frame.convertTo(detectedFrame, CV_8U);
		video.write(detectedFrame);

		imshow(kWinName, frame);
	}

	cap.release();
	video.release();
}

void mask_rcnn::postprocess(cv::Mat& frame, const std::vector<cv::Mat>& outs)
{
	Mat outDetections = outs[0];
	Mat outMasks = outs[1];

	//size[i]：The number of elements in each dimension
	const int numDetections = outDetections.size[2];
	const int numClasses = outMasks.size[1];

	/* Mat reshape(int cn, int rows = 0) const;
	*  @param cn New number of channels. If the parameter is 0, the number of channels remains the same.
	*  @param rows New number of rows.If the parameter is 0, the number of rows remains the same.
	*/
	outDetections = outDetections.reshape(1, outDetections.total() / 7);


	for (int i = 0; i < numDetections; ++i)
	{
		float score = outDetections.at<float>(i, 2);
		if (score > mfConfThreshold)
		{
			// Extract the bounding box
			int classId = static_cast<int>(outDetections.at<float>(i, 1));
			int left = static_cast<int>(frame.cols * outDetections.at<float>(i, 3));
			int top = static_cast<int>(frame.rows * outDetections.at<float>(i, 4));
			int right = static_cast<int>(frame.cols * outDetections.at<float>(i, 5));
			int bottom = static_cast<int>(frame.rows * outDetections.at<float>(i, 6));

			left = max(0, min(left, frame.cols - 1));
			top = max(0, min(top, frame.rows - 1));
			right = max(0, min(right, frame.cols - 1));
			bottom = max(0, min(bottom, frame.rows - 1));
			Rect box = Rect(left, top, right - left + 1, bottom - top + 1);

			// Extract the mask for the object, 
			Mat objectMask(outMasks.size[2], outMasks.size[3], CV_32F, outMasks.ptr<float>(i, classId));

			// Draw bounding box, colorize and show the mask on the image
			drawBox(frame, classId, score, box, objectMask);

		}
	}
}

void mask_rcnn::drawBox(cv::Mat& frame, int classId, float conf, cv::Rect box, cv::Mat& objectMask)
{
	//Draw a rectangle displaying the bounding box
	rectangle(frame, Point(box.x, box.y), Point(box.x + box.width, box.y + box.height), Scalar(255, 178, 50), 3);

	//Get the label for the class name and its confidence
	string label = format("%.2f", conf);
	if (!mvClasses.empty())
	{
		CV_Assert(classId < (int)mvClasses.size());
		label = mvClasses[classId] + ":" + label;
	}

	//Display the label at the top of the bounding box
	int baseLine;
	Size labelSize = getTextSize(label, FONT_HERSHEY_SIMPLEX, 0.5, 1, &baseLine);
	box.y = max(box.y, labelSize.height);
	rectangle(frame, Point(box.x, box.y - round(1.5*labelSize.height)), Point(box.x + round(1.5*labelSize.width), box.y + baseLine), Scalar(255, 255, 255), FILLED);
	putText(frame, label, Point(box.x, box.y), FONT_HERSHEY_SIMPLEX, 0.75, Scalar(0, 0, 0), 1);

	Scalar color = mvColors[classId%mvColors.size()];

	// Resize the mask, threshold, color and apply it on the image
	resize(objectMask, objectMask, Size(box.width, box.height));
	Mat mask = (objectMask > mfMaskThreshold);
	Mat coloredRoi = (0.3 * color + 0.7 * frame(box));
	coloredRoi.convertTo(coloredRoi, CV_8UC3);

	// Draw the contours on the image
	vector<Mat> contours;
	Mat hierarchy;

	mask.convertTo(mask, CV_8U);
	findContours(mask, contours, hierarchy, RETR_CCOMP, CHAIN_APPROX_SIMPLE);
	drawContours(coloredRoi, contours, -1, color, 5, LINE_8, hierarchy, 100);
	coloredRoi.copyTo(frame(box), mask);
}

main.cpp:

#include <string>
#include "mask_rcnn.h"

int main(int argc, char** argv)
{
	// Give the classes and colors files for the model
	string classesFile = "../model//mscoco_labels.names";
	string colorsFile = "../model//colors.txt";

	// Give the configuration and weight files for the model
	String modelWeights = "../model//frozen_inference_graph.pb";
	String textGraph = "../model//mask_rcnn_inception_v2_coco_2018_01_28.pbtxt";

	// Enter an image or video
	string imagepath = "../data//cars.jpg";
	string videopath = "../data//cars.mp4";

	// Output path settings
	std::string image_outputFile = "../result//mask_rcnn.jpg";
	std::string video_outputFile = "../result//mask_rcnn_out.avi";

	// Confidence threshold and Mask threshold
	mask_rcnn Mask_RCNN(0.5, 0.3);

	Mask_RCNN.detect_image(classesFile, colorsFile, textGraph, modelWeights, imagepath, image_outputFile);
	//Mask_RCNN.detect_video(classesFile, colorsFile, textGraph, modelWeights, videopath, video_outputFile);

	cv::waitKey(0);
	return 0;
}

3、實驗結果圖

效果還可以，就是僅使用CPU速度太慢了。此外，即是使用GPU速度也不是很快，，，，，，