PCA人脸识别

  在工作目录下建立/pic文件夹放入测试图片，建立/positive文件夹，放入自己的训练数据（我使用的是自己的相片中截获好的头像）建立list.txt，在里面写入pic/文件名以回车隔开，即可。

#define CV_NO_BACKWARD_COMPATIBILITY

#include "cv.h"
#include "highgui.h"

#include <iostream>
#include <cstdio>
#include <string.h>

#ifdef _EiC
#define WIN32
#endif

using namespace std;
using namespace cv;

void detectAndDraw( Mat& img,
                   CascadeClassifier& cascade, CascadeClassifier& nestedCascade,
                   double scale,int picid);

int compressPCA(const Mat& pcaset, int maxComponents,
const Mat& testset, Mat& compressed);

String cascadeName ="haarcascade_frontalface_alt.xml";
String nestedCascadeName ="haarcascade_eye_tree_eyeglasses.xml";

int WINDOWID=1;

int main( int argc, const char** argv )
{

    Mat frame, frameCopy, image;
    const String scaleOpt = "--scale=";
    size_t scaleOptLen = scaleOpt.length();
    const String cascadeOpt = "--cascade=";
    size_t cascadeOptLen = cascadeOpt.length();
    const String nestedCascadeOpt = "--nested-cascade";
    size_t nestedCascadeOptLen = nestedCascadeOpt.length();
    String inputName;

    CascadeClassifier cascade, nestedCascade;
    double scale = 1;

    for( int i = 1; i < argc; i++ )
    {
        if( cascadeOpt.compare( 0, cascadeOptLen, argv[i], cascadeOptLen ) == 0 )
            cascadeName.assign( argv[i] + cascadeOptLen );
        else if( nestedCascadeOpt.compare( 0, nestedCascadeOptLen, argv[i], nestedCascadeOptLen ) == 0 )
        {
            if( argv[i][nestedCascadeOpt.length()] == '=' )
                nestedCascadeName.assign( argv[i] + nestedCascadeOpt.length() + 1 );
            if( !nestedCascade.load( nestedCascadeName ) )
                cerr << "WARNING: Could not load classifier cascade for nested objects" << endl;
        }
        else if( scaleOpt.compare( 0, scaleOptLen, argv[i], scaleOptLen ) == 0 )
        {
            if( !sscanf( argv[i] + scaleOpt.length(), "%lf", &scale ) || scale < 1 )
                scale = 1;
        }
        else if( argv[i][0] == '-' )
        {
            cerr << "WARNING: Unknown option %s" << argv[i] << endl;
        }
        else
            inputName.assign( argv[i] );
    }

    if( !cascade.load( cascadeName ) )
    {
        cerr << "ERROR: Could not load classifier cascade" << endl;
        cerr << "Usage: facedetect [--cascade=\"<cascade_path>\"]\n"
            "   [--nested-cascade[=\"nested_cascade_path\"]]\n"
            "   [--scale[=<image scale>\n"
            "   [filename|camera_index]\n" ;
        return -1;
    }

    image = imread( inputName, 1 );

    if( !image.empty() )
    {
        detectAndDraw( image, cascade, nestedCascade, scale,0 );
        waitKey(0);
    }
    else if( !inputName.empty() )
    {
        FILE* f = fopen( inputName.c_str(), "rt" );
        if( f )
        {
            char buf[1000+1];
    int picid=1;
    printf("################################################################\n");
    printf("\n");
            while( fgets( buf, 1000, f ) )
            {

                int len = (int)strlen(buf), c;
                while( len > 0 && isspace(buf[len-1]) )
                len--;
                buf[len] = '\0';
                cout << "file " << buf << endl;
                image = imread( buf, 1 );
                if( !image.empty() )
                {
                    detectAndDraw( image, cascade, nestedCascade, scale,picid );
      picid++;
      printf("\n");
      printf("################################################################\n");
      printf("\n");
      printf("图片编号: %d \n",picid);
                 }
     
             }
             fclose(f);
    waitKey(0);
        }
    }

    return 0;
}

void detectAndDraw( Mat& img,
                   CascadeClassifier& cascade, CascadeClassifier& nestedCascade,
                   double scale, int picid)
{
    int i = 0;
    double t = 0;
    vector<Rect> faces;
    const static Scalar colors[] = { CV_RGB(0,0,255),
        CV_RGB(0,128,255),
        CV_RGB(0,255,255),
        CV_RGB(0,255,0),
        CV_RGB(255,128,0),
        CV_RGB(255,255,0),
        CV_RGB(255,0,0),
        CV_RGB(255,0,255)} ;
    Mat gray, smallImg( cvRound (img.rows/scale), cvRound(img.cols/scale), CV_8UC1 );

    cvtColor( img, gray, CV_BGR2GRAY );
    resize( gray, smallImg, smallImg.size(), 0, 0, INTER_LINEAR );
    equalizeHist( smallImg, smallImg );

    t = (double)cvGetTickCount();
    cascade.detectMultiScale( smallImg, faces,
        1.1, 2, 0
        |CV_HAAR_SCALE_IMAGE
        ,
        Size(30, 30) );
    t = (double)cvGetTickCount() - t;
    printf( "检测用时为： %g ms\n", t/((double)cvGetTickFrequency()*1000.) );

int index=1;
char facenum[100];
char filenum[100];
char filename[255];
//pca算法处理图像开始

Mat p1,p2,p3,p4,compressed,UnknowFaceMat;
p1=cv::imread("positive/1.jpg",0);
p2=cv::imread("positive/2.jpg",0);
p3=cv::imread("positive/3.jpg",0);
p4=cv::imread("positive/4.jpg",0);
float dataArr[40000];
int start=0;
for(int i=0;i<p1.rows;i++)
{
   for(int j=0;j<p1.cols;j++)
   {
    dataArr[start]=(float) p1.at<uchar>(i,j);
    start++;
   }
}
for(int i=0;i<p2.rows;i++)
{
   for(int j=0;j<p2.cols;j++)
   {
    dataArr[start]=(float)p2.at<uchar>(i,j);
    start++;
   }
}
for(int i=0;i<p3.rows;i++)
{
   for(int j=0;j<p3.cols;j++)
   {
    dataArr[start]=(float)p3.at<uchar>(i,j);
    start++;
   }
}
for(int i=0;i<p4.rows;i++)
{
   for(int j=0;j<p4.cols;j++)
   {
    dataArr[start]=(float)p4.at<uchar>(i,j);
    start++;
   }
}
Mat positiveMat(4,10000,CV_32FC1,dataArr);
//pca算法结束

    for( vector<Rect>::const_iterator r = faces.begin(); r != faces.end(); r++, i++ )
    {
        Mat smallImgROI,tempMat;
        vector<Rect> nestedObjects;
        Point center;
        Scalar color = colors[i%8];
        int radius;
        center.x = cvRound((r->x + r->width*0.5)*scale);
        center.y = cvRound((r->y + r->height*0.5)*scale);
        radius = cvRound((r->width + r->height)*0.25*scale);
       // circle( img, center, radius, color, 3, 8, 0 );
      
        smallImgROI = smallImg(*r);

   tempMat = smallImg(*r);
  
   itoa(picid,filenum,10);
   itoa(index,facenum,10);
   strcpy(filename,"faces/");
   strcat(filename,filenum);
   strcat(filename,"_");
   strcat(filename,facenum);
   strcat(filename,".jpg");
  
   cv::Mat graysmallface(Size(100,100),CV_8UC1);
   cv::resize(tempMat,graysmallface,graysmallface.size(),0,0);

  
   cv::imwrite(filename,graysmallface);
   printf("检测到人脸，写入如下文件: %s\n", filename);

   index++;
  
   UnknowFaceMat=graysmallface;
   float UnknowDataArr[10000];
   start=0;
   for(int i=0;i<UnknowFaceMat.rows;i++)
   {
    for(int j=0;j<UnknowFaceMat.cols;j++)
    {
     UnknowDataArr[start]=(float) UnknowFaceMat.at<uchar>(i,j);
     start++;
    }
   }

   Mat negativeMat(1,10000,CV_32FC1,UnknowDataArr);

   int result=compressPCA(positiveMat,3,negativeMat,compressed);
   if(result==-1)
   {
    continue;
   }

   //--use pca end---

   if(result==1)
   {  
    printf("检测到指定的人！\n点任意键继续\n");

    center.x = cvRound((r->x + r->width*0.5)*scale);
    center.y = cvRound((r->y + r->height*0.5)*scale);
    radius = cvRound((r->width + r->height)*0.25*scale);
    circle( img, center, radius, color, 2, 8, 0 );
   
    char windowname[255];
    char wnum[10];
    itoa(WINDOWID,wnum,10);
    strcpy(windowname,"Match_");
    strcat(windowname,wnum);
  
    cv::Mat tempimg(Size(800,600),CV_8UC3);
    cv::resize(img,tempimg,tempimg.size(),0,0);
   
    cv::imshow(windowname,img);
    WINDOWID++;
    waitKey(0);
   }
    }
}

int compressPCA(const Mat& pcaset, int maxComponents,
const Mat& testset, Mat& compressed)
{
PCA pca(pcaset, // pass the data
Mat(), // we do not have a pre-computed mean vector,
// so let the PCA engine to compute it
CV_PCA_DATA_AS_ROW, // indicate that the vectors
// are stored as matrix rows
// (use CV_PCA_DATA_AS_COL if the vectors are
// the matrix columns)
maxComponents // specify, how many principal components to retain
);
// if there is no test data, just return the computed basis, ready-to-use
//if( !testset.data )
//return pca;
CV_Assert( testset.cols == pcaset.cols );
compressed.create(testset.rows, maxComponents, testset.type());
Mat reconstructed;
for( int i = 0; i < testset.rows; i++ )
{
Mat vec = testset.row(i), coeffs = compressed.row(i);
// compress the vector, the result will be stored
// in the i-th row of the output matrix
pca.project(vec, coeffs);
// and then reconstruct it
pca.backProject(coeffs, reconstructed);
// and measure the error
printf("%d. 偏移量为： %g\n", i+1, norm(vec, reconstructed, NORM_L2));

if(norm(vec, reconstructed, NORM_L2)>2000)
{
   return -1;
}

}
return 1;
}