intraFace人脸对齐

/////////////////////////////////////////////////////////////////////////////////////////////////////// /// DemoTracker.cpp /// /// Description: /// This program shows you how to use FaceAlignment class in tracking facial landmarks in a video or realtime. /// There are two modes: VIDEO, REALTIME. /// In the VIDEO mode, the program reads input from a video and perform tracking. /// In the REALTIME mode, the program reads input from the first camera it finds and perform tracking. /// Note that tracking is performed on the largest face found. The face is detected through OpenCV. /// In this version, we add head pose estimation. /// /// Dependencies: None. OpenCV DLLs and include folders are copied. /// /// Author: Xuehan Xiong, xiong828@gmail.com /// /// Creation Date: 1/24/2014 /// /// Version: 1.2 /// /// Citation: /// Xuehan Xiong, Fernando de la Torre, Supervised Descent Method and Its Application to Face Alignment. CVPR, 2013 /////////////////////////////////////////////////////////////////////////////////////////////////////// #include <opencv2/core/core.hpp> #include <opencv2/objdetect/objdetect.hpp> #include <opencv2/highgui/highgui.hpp> #include <opencv2/imgproc/imgproc.hpp> #include <opencv2/opencv.hpp> #include <iostream> #include <algorithm> #include <string> #include <vector> #include <time.h> #include <Windows.h> #include <intraface/FaceAlignment.h> #include <intraface/XXDescriptor.h> using namespace std; using namespace cv; bool compareRect(cv::Rect r1, cv::Rect r2) { return r1.height < r2.height; } // 2 modes: REALTIME,VIDEO #define REALTIME void drawPose(cv::Mat& img, const cv::Mat& rot, float lineL) { int loc[2] = {70, 70}; int thickness = 2; int lineType = 8; cv::Mat P = (cv::Mat_<float>(3,4) << 0, lineL, 0, 0, 0, 0, -lineL, 0, 0, 0, 0, -lineL); P = rot.rowRange(0,2)*P; P.row(0) += loc[0]; P.row(1) += loc[1]; cv::Point p0(P.at<float>(0,0),P.at<float>(1,0)); line(img, p0, cv::Point(P.at<float>(0,1),P.at<float>(1,1)), cv::Scalar( 255, 0, 0 ), thickness, lineType); line(img, p0, cv::Point(P.at<float>(0,2),P.at<float>(1,2)), cv::Scalar( 0, 255, 0 ), thickness, lineType); line(img, p0, cv::Point(P.at<float>(0,3),P.at<float>(1,3)), cv::Scalar( 0, 0, 255 ), thickness, lineType); //printf("%f %f %f\n", rot.at<float>(0, 0), rot.at<float>(0, 1), rot.at<float>(0, 2)); //printf("%f %f %f\n", rot.at<float>(1, 0), rot.at<float>(1, 1), rot.at<float>(1, 2)); //printf("%f %f %f\n", rot.at<float>(2, 0), rot.at<float>(2, 1), rot.at<float>(2, 2)); Vec3d eav; Mat tmp,tmp1,tmp2,tmp3,tmp4,tmp5; double _pm[12] = {rot.at<float>(0, 0), rot.at<float>(0, 1),rot.at<float>(0, 2), 0, rot.at<float>(1, 0), rot.at<float>(1, 1),rot.at<float>(1, 2),0, rot.at<float>(2, 0),rot.at<float>(2, 1),rot.at<float>(2, 2),0}; decomposeProjectionMatrix(Mat(3,4,CV_64FC1,_pm),tmp,tmp1,tmp2,tmp3,tmp4,tmp5,eav); stringstream ss; ss << eav[0]; string txt = "Pitch: " + ss.str(); putText(img, txt, Point(60, 20), 0.5,0.5,Scalar(255,255,255)); stringstream ss1; ss1 << eav[1]; string txt1 = "Yaw: " + ss1.str(); putText(img, txt1, Point(60, 40), 0.5,0.5,Scalar(255,255,255)); stringstream ss2; ss2 << eav[2]; string txt2 = "Roll: " + ss2.str(); putText(img, txt2, Point(60, 60), 0.5,0.5,Scalar(255,255,255)); } int main() { char detectionModel[] = "../models/DetectionModel-v1.5.bin"; char trackingModel[] = "../models/DetectionModel-v1.5.bin"; string faceDetectionModel("../models/haarcascade_frontalface_alt2.xml"); // initialize a XXDescriptor object INTRAFACE::XXDescriptor xxd(4); // initialize a FaceAlignment object INTRAFACE::FaceAlignment fa(detectionModel, detectionModel, &xxd); if (!fa.Initialized()) { cerr << "FaceAlignment cannot be initialized." << endl; return -1; } // load OpenCV face detector model cv::CascadeClassifier face_cascade; if( !face_cascade.load( faceDetectionModel ) ) { cerr << "Error loading face detection model." << endl; return -1; } #ifdef REALTIME // use the first camera it finds cv::VideoCapture cap(0); #endif #ifdef VIDEO string filename("../data/vid.wmv"); cv::VideoCapture cap(filename); #endif if(!cap.isOpened()) return -1; int key = 0; bool isDetect = true; bool eof = false; float score, notFace = 0.3; cv::Mat X,X0; string winname("Demo IntraFace Tracker"); cv::namedWindow(winname); while (key!=27) // Press Esc to quit { cv::Mat frame; cap >> frame; // get a new frame from camera //frame = cv::imread("xjp4.jpg"); if (frame.rows == 0 || frame.cols == 0) break; if (isDetect) { // face detection vector<cv::Rect> faces; face_cascade.detectMultiScale(frame, faces, 1.2, 2, 0, cv::Size(50, 50)); // if no face found, do nothing if (faces.empty()) { cv::imshow(winname,frame); key = cv::waitKey(5); continue ; } clock_t start_time1 = clock(); // facial feature detection on largest face found if (fa.Detect(frame,*max_element(faces.begin(),faces.end(),compareRect),X0,score) != INTRAFACE::IF_OK) break; clock_t finish_time1 = clock(); double total_time = (double)(finish_time1-start_time1)/CLOCKS_PER_SEC; std::cout<< "detect time: " << total_time*1000 << endl; isDetect = false; } else { clock_t start_time1 = clock(); // facial feature tracking if (fa.Track(frame,X0,X,score) != INTRAFACE::IF_OK) break; clock_t finish_time1 = clock(); double total_time = (double)(finish_time1-start_time1)/CLOCKS_PER_SEC; std::cout<< "track time: " << total_time*1000 << endl; X0 = X; } if (score < notFace) // detected face is not reliable isDetect = true; else { // plot facial landmarks for (int i = 0 ; i < X0.cols ; i++) cv::circle(frame,cv::Point((int)X0.at<float>(0,i), (int)X0.at<float>(1,i)), 1, cv::Scalar(0,255,0), -1); // head pose estimation INTRAFACE::HeadPose hp; fa.EstimateHeadPose(X0,hp); // plot head pose drawPose(frame, hp.rot, 50); } printf("score %f\n", score); cv::imshow(winname,frame); key = cv::waitKey(5); } return 0; }