本程序首先利用从摄像头检测到的人脸图片，先进行直方图均衡化并缩放到92*112的图片大小，然后根据train.txt的采集到的人脸模版进行匹配识别（最好是在统一光照下，采集不同角度的人脸图片各一张）注意：影响的极大因素在于光照，模版若与采集的图像光照不一样，识别率很低。经测试，模板若与检测的图像在同一光照下的话，侧脸，仰脸，正脸均可识别，且识别率较高
//#include <stdio.h>#include <string.h>#include "cv.h"#include "cvaux.h"#include "highgui.h"#include <stdlib.h>#include <assert.h> #include <math.h> #include <float.h> #include <limits.h> #include <time.h> #include <ctype.h> ////定义几个重要的全局变量IplImage ** faceImgArr = 0; // 指向训练人脸和测试人脸的指针（在学习和识别阶段指向不同）CvMat * personNumTruthMat = 0; // 人脸图像的ID号int nTrainFaces = 0; // 训练图像的数目int nEigens = 0; // 自己取的主要特征值数目IplImage * pAvgTrainImg = 0; // 训练人脸数据的平均值IplImage ** eigenVectArr = 0; // 投影矩阵，也即主特征向量CvMat * eigenValMat = 0; // 特征值CvMat * projectedTrainFaceMat = 0; // 训练图像的投影char *filename[5]={"face1.jpg","face2.jpg","face3.jpg","face4.jpg","face5.jpg"};static CvMemStorage* storage = 0; static CvHaarClassifierCascade* cascade = 0; int j=0;//统计记录的人脸数char a[512]={0};int a1,a2,a3,a4;time_t timeBegin, timeEnd; int timeuse; //// 函数原型void learn();void doPCA();void storeTrainingData();int loadTrainingData(CvMat ** pTrainPersonNumMat);int findNearestNeighbor(float * projectedTestFace);int loadFaceImgArray(char * filename);void printUsage();int detect_and_draw( IplImage* image ); int recognize(IplImage *faceimg);//主函数，主要包括学习和识别两个阶段，需要运行两次，通过命令行传入的参数区分int main( int argc, char** argv ){ CvCapture* capture = 0; IplImage *frame, *frame_copy = 0; int optlen = strlen("--cascade="); char *cascade_name = "H://opencv-2.4.3//opencv//data//haarcascades//haarcascade_frontalface_alt2.xml"; //opencv装好后haarcascade_frontalface_alt2.xml的路径, //也可以把这个文件拷到你的工程文件夹下然后不用写路径名cascade_name= "haarcascade_frontalface_alt2.xml"; //或者cascade_name ="C:\\Program Files\\OpenCV\\data\\haarcascades\\haarcascade_frontalface_alt2.xml" cascade = (CvHaarClassifierCascade*)cvLoad( cascade_name, 0, 0, 0 ); if( !cascade ) { fprintf( stderr, "ERROR: Could not load classifier cascade\n" ); fprintf( stderr, "Usage: facedetect --cascade=\"<cascade_path>\" [filename|camera_index]\n" ); return -1; } storage = cvCreateMemStorage(0); capture = cvCreateCameraCapture(-1); cvNamedWindow( "result", 1 ); if( capture ) { timeBegin = time(NULL); learn(); for(;;) { timeEnd = time(NULL); timeuse=timeEnd - timeBegin;//计算经过的时间,统计人数 if( !cvGrabFrame( capture )) break; frame = cvRetrieveFrame( capture ); if( !frame ) break; if( !frame_copy ) frame_copy = cvCreateImage( cvSize(frame->width,frame->height), IPL_DEPTH_8U, frame->nChannels ); if( frame->origin == IPL_ORIGIN_TL )//如果图像的起点在左上角 cvCopy( frame, frame_copy, 0 ); else cvFlip( frame, frame_copy, 0 );//如果图像的起点不在左上角，而在左下角时，进行X轴对称 detect_and_draw( frame_copy ); //检测并且识别 if( cvWaitKey( 10 ) >= 0 ) break; } cvReleaseImage( &frame_copy ); cvReleaseCapture( &capture ); } else { printf("Cannot read from CAM"); return -1; } cvDestroyWindow("result"); return 0; } //学习阶段代码void learn(){int i, offset;//加载训练图像集nTrainFaces = loadFaceImgArray("train.txt");if( nTrainFaces < 2 ){fprintf(stderr,"Need 2 or more training faces\n""Input file contains only %d\n", nTrainFaces);return;}// 进行主成分分析doPCA();//将训练图集投影到子空间中projectedTrainFaceMat = cvCreateMat( nTrainFaces, nEigens, CV_32FC1 );offset = projectedTrainFaceMat->step / sizeof(float);for(i=0; i<nTrainFaces; i++){//int offset = i * nEigens;cvEigenDecomposite(faceImgArr[i],nEigens,eigenVectArr,0, 0,pAvgTrainImg,//projectedTrainFaceMat->data.fl + i*nEigens);projectedTrainFaceMat->data.fl + i*offset);}//将训练阶段得到的特征值，投影矩阵等数据存为.xml文件，以备测试时使用storeTrainingData();}//加载保存过的训练结果int loadTrainingData(CvMat ** pTrainPersonNumMat){CvFileStorage * fileStorage;int i;fileStorage = cvOpenFileStorage( "facedata.xml", 0, CV_STORAGE_READ );if( !fileStorage ){fprintf(stderr, "Can't open facedata.xml\n");return 0;}nEigens = cvReadIntByName(fileStorage, 0, "nEigens", 0);nTrainFaces = cvReadIntByName(fileStorage, 0, "nTrainFaces", 0);*pTrainPersonNumMat = (CvMat *)cvReadByName(fileStorage, 0, "trainPersonNumMat", 0);eigenValMat = (CvMat *)cvReadByName(fileStorage, 0, "eigenValMat", 0);projectedTrainFaceMat = (CvMat *)cvReadByName(fileStorage, 0, "projectedTrainFaceMat", 0);pAvgTrainImg = (IplImage *)cvReadByName(fileStorage, 0, "avgTrainImg", 0);eigenVectArr = (IplImage **)cvAlloc(nTrainFaces*sizeof(IplImage *));for(i=0; i<nEigens; i++){char varname[200];sprintf( varname, "eigenVect_%d", i );eigenVectArr[i] = (IplImage *)cvReadByName(fileStorage, 0, varname, 0);}cvReleaseFileStorage( &fileStorage );return 1;}//存储训练结果void storeTrainingData(){CvFileStorage * fileStorage;int i;fileStorage = cvOpenFileStorage( "facedata.xml", 0, CV_STORAGE_WRITE );//存储特征值，投影矩阵，平均矩阵等训练结果cvWriteInt( fileStorage, "nEigens", nEigens );cvWriteInt( fileStorage, "nTrainFaces", nTrainFaces );cvWrite(fileStorage, "trainPersonNumMat", personNumTruthMat, cvAttrList(0,0));cvWrite(fileStorage, "eigenValMat", eigenValMat, cvAttrList(0,0));cvWrite(fileStorage, "projectedTrainFaceMat", projectedTrainFaceMat, cvAttrList(0,0));cvWrite(fileStorage, "avgTrainImg", pAvgTrainImg, cvAttrList(0,0));for(i=0; i<nEigens; i++){char varname[200];sprintf( varname, "eigenVect_%d", i );cvWrite(fileStorage, varname, eigenVectArr[i], cvAttrList(0,0));}cvReleaseFileStorage( &fileStorage );}//寻找最接近的图像int findNearestNeighbor(float * projectedTestFace){double leastDistSq = DBL_MAX;//定义最小距离，并初始化为无穷大int i, iTrain, iNearest = 0;for(iTrain=0; iTrain<nTrainFaces; iTrain++){double distSq=0;for(i=0; i<nEigens; i++){float d_i =projectedTestFace[i] -projectedTrainFaceMat->data.fl[iTrain*nEigens + i];distSq += d_i*d_i / eigenValMat->data.fl[i]; // Mahalanobis算法计算的距离，差的距离的平方除以平均脸的特征值//distSq += d_i*d_i; // Euclidean算法计算的距离}if(distSq < leastDistSq){leastDistSq = distSq;iNearest = iTrain;}}//printf("leastdistsq==%f",leastDistSq);return iNearest;}//主成分分析void doPCA(){int i;CvTermCriteria calcLimit;CvSize faceImgSize;// 自己设置主特征值个数nEigens = nTrainFaces-1;//分配特征向量存储空间faceImgSize.width = faceImgArr[0]->width;faceImgSize.height = faceImgArr[0]->height;eigenVectArr = (IplImage**)cvAlloc(sizeof(IplImage*) * nEigens);//分配个数为住特征值个数for(i=0; i<nEigens; i++)eigenVectArr[i] = cvCreateImage(faceImgSize, IPL_DEPTH_32F, 1);//分配主特征值存储空间eigenValMat = cvCreateMat( 1, nEigens, CV_32FC1 );// 分配平均图像存储空间pAvgTrainImg = cvCreateImage(faceImgSize, IPL_DEPTH_32F, 1);// 设定PCA分析结束条件calcLimit = cvTermCriteria( CV_TERMCRIT_ITER, nEigens, 1);//最大迭代次数为nEigens// 计算平均图像，特征值，特征向量cvCalcEigenObjects(nTrainFaces,(void*)faceImgArr,(void*)eigenVectArr,CV_EIGOBJ_NO_CALLBACK,0,0,&calcLimit,pAvgTrainImg,eigenValMat->data.fl//存储求得的eigenvalue);cvNormalize(eigenValMat, eigenValMat, 1, 0, CV_L1, 0);}//加载txt文件的列举的图像int loadFaceImgArray(char * filename){FILE * imgListFile = 0;char imgFilename[512];int iFace, nFaces=0;if( !(imgListFile = fopen(filename, "r")) ){fprintf(stderr, "Can\'t open file %s\n", filename);return 0;}// 统计人脸数while( fgets(imgFilename, 512, imgListFile) ) ++nFaces;//char *fgets(char *buf, int bufsize, FILE *stream);从文件结构体指针stream中读取数据，每次读取一行。读取的数据保存在buf指向的字符数组中，每次最多读取bufsize-1个字符（第bufsize个字符赋'\0'），如果文件中的该行，不足bufsize个字符，则读完该行就结束。如果函数读取成功，则返回指针buf，失败则返回NULL。rewind(imgListFile);//将文件内部的位置指针重新指向一个流（数据流/文件）的开头// 分配人脸图像存储空间和人脸ID号存储空间faceImgArr = (IplImage **)cvAlloc( nFaces*sizeof(IplImage *) );personNumTruthMat = cvCreateMat( 1, nFaces, CV_32SC1 );//CvMat* cvCreateMat( int rows, int cols, int type );for(iFace=0; iFace<nFaces; iFace++){// 从文件中读取序号和人脸名称fscanf(imgListFile,"%d %s", personNumTruthMat->data.i+iFace, imgFilename);// fscanf(FILE *stream, char *format,[argument...])功 能: 从一个流中执行格式化输入,fscanf遇到空格和换行时结束// 加载人脸图像faceImgArr[iFace] = cvLoadImage(imgFilename, CV_LOAD_IMAGE_GRAYSCALE);if( !faceImgArr[iFace] ){fprintf(stderr, "Can\'t load image from %s\n", imgFilename);return 0;}}fclose(imgListFile);return nFaces;}//void printUsage(){printf("Usage: eigenface <command>\n"," Valid commands are\n"" train\n"" test\n");}int detect_and_draw( IplImage* img ) { CvFont font; cvInitFont( &font, CV_FONT_VECTOR0,1, 1, 0, 1, 8); static CvScalar colors[] = { {{0,0,255}}, {{0,128,255}}, {{0,255,255}}, {{0,255,0}}, {{255,128,0}}, {{255,255,0}}, {{255,0,0}}, {{255,0,255}} }; double scale = 1.3; IplImage* gray = cvCreateImage( cvSize(img->width,img->height), 8, 1 ); IplImage* small_img = cvCreateImage( cvSize( cvRound (img->width/scale), cvRound (img->height/scale)), 8, 1 ); int i,personnum=0; cvCvtColor( img, gray, CV_BGR2GRAY ); cvResize( gray, small_img, CV_INTER_LINEAR ); cvEqualizeHist( small_img, small_img ); cvClearMemStorage( storage ); if( cascade ) { double t = (double)cvGetTickCount(); CvSeq* faces = cvHaarDetectObjects( small_img, cascade, storage, 1.1, 2, 0/*CV_HAAR_DO_CANNY_PRUNING*/, cvSize(30, 30) ); t = (double)cvGetTickCount() - t; // printf( "detection time = %gms\n", t/((double)cvGetTickFrequency()*1000.) ); IplImage* temp_img=cvCreateImage(cvSize(92,112),8,1); for( i = 0; i < (faces ? faces->total : 0); i++ ) { CvRect* r = (CvRect*)cvGetSeqElem( faces, i ); IplImage *dst=cvCreateImage(cvSize(r->width,r->height),8,1);//cvsize只能选取r->width,r->height不能再后面*scale或+100 CvPoint p1; p1.x=cvRound((r->x)*scale); p1.y=cvRound((r->y)*scale); CvPoint p2; p2.x=cvRound((r->x+r->width)*scale); p2.y=cvRound((r->y+r->height)*scale); cvRectangle(img,p1,p2,colors[i%8],3,8,0); cvSetImageROI(small_img,*r); cvCopy(small_img,dst); cvResize(dst,temp_img); cvEqualizeHist( temp_img, temp_img ); cvResetImageROI(small_img); cvSaveImage(filename[i],temp_img); cvReleaseImage(&dst); //开始识别temp_img personnum=recognize(temp_img); if(personnum==1) cvPutText(img, "Yanming" , cvPoint(20, 20), &font, CV_RGB(255,255,255));//将正确识别的人的姓名显示在屏幕上 } } cvShowImage( "result", img ); cvReleaseImage( &gray ); cvReleaseImage( &small_img ); return -1;} int recognize(IplImage *faceimg){int i, nTestFaces = 0; // 测试人脸数CvMat * trainPersonNumMat = 0; // 训练阶段的人脸数float * projectedTestFace = 0;// 加载保存在.xml文件中的训练结果if( !loadTrainingData( &trainPersonNumMat ) ) return -3;projectedTestFace = (float *)cvAlloc( nEigens*sizeof(float) );int iNearest, nearest;//将测试图像投影到子空间中cvEigenDecomposite(faceimg,nEigens,eigenVectArr,0, 0,pAvgTrainImg,projectedTestFace);//cvNormalize(projectedTestFace, projectedTestFace, 1, 0, CV_L1, 0);iNearest = findNearestNeighbor(projectedTestFace);nearest = trainPersonNumMat->data.i[iNearest];printf("nearest = %d", nearest);if(timeuse<=10){if((nearest==1)|(nearest==11)|(nearest==111))//可以更改train.txt中训练图片的编号,这里将侧脸，仰脸，正脸都归为一起a1++;if(nearest==2)a2++;if(nearest==3)a3++;if(nearest==4)a4++;if(a1>7)//如果10s中识别的次数为6则认定为a1{printf("yanming\n");return 1;}if(a2>6){printf("others\n");}if(a3>6){printf("ma\n");}if(a4>6){printf("ba\n");}}else{timeBegin=time(NULL);a1=0;a2=0;a3=0;a4=0;return 0;}return -1;}
