opencv联合dlib视频人脸识别例子

#include <dlib/image_processing/frontal_face_detector.h> #include <dlib/image_processing/render_face_detections.h> #include <dlib/image_processing.h> #include <dlib/gui_widgets.h> #include <dlib/image_io.h> #include <dlib/opencv.h> #include <dlib/dnn.h> #include <dlib/data_io.h> #include <dlib/clustering.h> #include <dlib/string.h> #include <opencv2/highgui/highgui.hpp> #include <opencv2/core/core.hpp> #include <opencv2/opencv.hpp> #include <opencv2/imgproc/imgproc.hpp> #include <iostream> #include <vector> #include <ctime> #include <string> #include <map> #include <sstream> #ifdef __cplusplus extern "C"{ #endif #include <stdlib.h> #include <string.h> #include <unistd.h> #include <dirent.h> #ifdef __cplusplus } #endif //由于dlib和opencv中有相当一部分类同名，故不能同时对它们使用using namespace，否则会出现一些莫名其妙的问题 //且dlib库和标准std库中的类发生冲突，如map,string 类等等 using namespace std; using namespace cv; //using namespace dlib; void getFiles(std::string path, std::map<std::string, std::string> &files); void line_one_face_detections(cv::Mat img, std::vector<dlib::full_object_detection> fs); //定义好一堆模板别名，以供后续方便使用 template <template <int,template<typename>class,int,typename> class block, int N, template<typename>class BN, typename SUBNET> using residual = dlib::add_prev1<block<N,BN,1,dlib::tag1<SUBNET>>>; template <template <int,template<typename>class,int,typename> class block, int N, template<typename>class BN, typename SUBNET> using residual_down = dlib::add_prev2<dlib::avg_pool<2,2,2,2,dlib::skip1<dlib::tag2<block<N,BN,2,dlib::tag1<SUBNET>>>>>>; template <int N, template <typename> class BN, int stride, typename SUBNET> using block = BN<dlib::con<N,3,3,1,1,dlib::relu<BN<dlib::con<N,3,3,stride,stride,SUBNET>>>>>; template <int N, typename SUBNET> using ares = dlib::relu<residual<block,N,dlib::affine,SUBNET>>; template <int N, typename SUBNET> using ares_down = dlib::relu<residual_down<block,N,dlib::affine,SUBNET>>; template <typename SUBNET> using alevel0 = ares_down<256,SUBNET>; template <typename SUBNET> using alevel1 = ares<256,ares<256,ares_down<256,SUBNET>>>; template <typename SUBNET> using alevel2 = ares<128,ares<128,ares_down<128,SUBNET>>>; template <typename SUBNET> using alevel3 = ares<64,ares<64,ares<64,ares_down<64,SUBNET>>>>; template <typename SUBNET> using alevel4 = ares<32,ares<32,ares<32,SUBNET>>>; using anet_type = dlib::loss_metric<dlib::fc_no_bias<128,dlib::avg_pool_everything< alevel0< alevel1< alevel2< alevel3< alevel4< dlib::max_pool<3,3,2,2,dlib::relu<dlib::affine<dlib::con<32,7,7,2,2, dlib::input_rgb_image_sized<150> >>>>>>>>>>>>; /* 识别视频中的某一帧图像中是不是有库里的某个人 方法： 统计出库文件夹中所有人的图片的face_descriptors，然后计算出当前图片中的人脸face_descriptors，二者之间距离小于0.6则视为同一个人 ./t11 hls_faces huanlesong.mp4 */ int main(int argc, char *argv[]) { time_t start_t, end_t; if(argc != 3) { std::cout<< "you should specified a dir and a video stream!"<<std::endl; return 0; } time(&start_t); std::map<string, string> files; getFiles(argv[1], files); if(files.empty()) { std::cout<< "No pic files found in "<< argv[1] <<std::endl; return 0; } //加载训练好的级联分类器，利用haar级联分类器快速找出人脸区域，然后交给dlib检测人脸部位 cv::CascadeClassifier faceDetector("haarcascade_frontalface_alt2.xml"); //cv::CascadeClassifier faceDetector("./output/cascade.xml"); if(faceDetector.empty()) { std::cout << "face detector is empty!" <<std::endl; return 0; } //加载人脸形状探测器 dlib::shape_predictor sp; dlib::deserialize("./shape_predictor_68_face_landmarks.dat") >> sp; //加载负责人脸识别的DNN anet_type net; dlib::deserialize("dlib_face_recognition_resnet_model_v1.dat") >> net; //人脸描述符库, face_descriptor ---> name map<dlib::matrix<float,0,1>, string> fdlib; for(map<string, string>::iterator it = files.begin(); it != files.end(); it++ ) { std::cout << "filename:" << it->second << " filepath:" <<it->first<<std::endl; cv::Mat frame = cv::imread(it->first); cv::Mat src; cv::cvtColor(frame, src, CV_BGR2GRAY); dlib::array2d<dlib::bgr_pixel> dimg; dlib::assign_image(dimg, dlib::cv_image<uchar>(src)); //haar级联分类器探测人脸区域，获取一系列人脸所在区域 std::vector<cv::Rect> objects; std::vector<dlib::rectangle> dets; faceDetector.detectMultiScale(src, objects); for (int i = 0; i < objects.size(); i++) { //cv::rectangle(frame, objects[i], CV_RGB(200,0,0)); dlib::rectangle r(objects[i].x, objects[i].y, objects[i].x + objects[i].width, objects[i].y + objects[i].height); dets.push_back(r); //正常情况下应该只检测到一副面容 } if (dets.size() == 0) continue; std::vector<dlib::matrix<dlib::rgb_pixel>> faces; std::vector<dlib::full_object_detection> shapes; for(int i = 0; i < dets.size(); i++) { dlib::full_object_detection shape = sp(dimg, dets[i]); //获取指定一个区域的人脸形状 shapes.push_back(shape); dlib::matrix<dlib::rgb_pixel> face_chip; dlib::extract_image_chip(dimg, dlib::get_face_chip_details(shape,150,0.25), face_chip); faces.push_back(move(face_chip)); } if (faces.size() == 0) { cout << "No faces found in " << it->second<<endl; continue; } std::vector<dlib::matrix<float,0,1>> face_descriptors = net(faces); for(std::vector<dlib::matrix<float,0,1>>::iterator iter = face_descriptors.begin(); iter != face_descriptors.end(); iter++ ) { fdlib.insert(pair<dlib::matrix<float,0,1>, string>(*iter, it->second)); } } time(&end_t); std::cout << "ok, all pic in lib had been keep on. use time:"<< end_t - start_t << " s" <<std::endl; //加载视频 VideoCapture capture(argv[2]); while(true) { //加载待检测的图片 cv::Mat frame; capture >> frame; if (frame.empty()) break; cv::Mat src; cv::cvtColor(frame, src, CV_BGR2GRAY); dlib::array2d<dlib::bgr_pixel> dimg; dlib::assign_image(dimg, dlib::cv_image<uchar>(src)); //haar级联分类器探测人脸区域，获取一系列人脸所在区域 std::vector<cv::Rect> objects; std::vector<dlib::rectangle> dets; faceDetector.detectMultiScale(src, objects); for (int i = 0; i < objects.size(); i++) { cv::rectangle(frame, objects[i], CV_RGB(200,0,0)); dlib::rectangle r(objects[i].x, objects[i].y, objects[i].x + objects[i].width, objects[i].y + objects[i].height); dets.push_back(r); //正常情况下应该只检测到一副面容 } if (dets.size() == 0) { continue; } std::vector<dlib::matrix<dlib::rgb_pixel>> faces; std::vector<dlib::full_object_detection> shapes; for(int i = 0; i < dets.size(); i++) { dlib::full_object_detection shape = sp(dimg, dets[i]); //获取指定一个区域的人脸形状 shapes.push_back(shape); dlib::matrix<dlib::rgb_pixel> face_chip; dlib::extract_image_chip(dimg, dlib::get_face_chip_details(shape,150,0.25), face_chip); faces.push_back(move(face_chip)); } if (faces.size() == 0) { continue; } line_one_face_detections(frame, shapes); std::vector<dlib::matrix<float,0,1>> face_descriptors = net(faces); //遍历库，查找相似图像 float min_distance = 0.7; std::string similar_name = "unknown"; for(map<dlib::matrix<float,0,1>, string>::iterator it=fdlib.begin(); it != fdlib.end(); it++ ) { float distance = length(it->first - face_descriptors[0]); if( distance < 0.5 ) //应该计算一个最近值 { if( distance <= min_distance) { mi