如何在OpenCV中检测多个面孔？

Question

I am having problem with detecting multiple faces at a time.the following code detect a frontal face and eyes for one person only.I tried to make it work for all available faces but i can't...please help

  CvCapture* capture;
    Mat frame;
    std::vector<Rect> faces;
    Mat frame_gray;
    Mat frame;
    CascadeClassifier face_cascade;
    CascadeClassifier eyes_cascade;

    face_cascade.load( "haarcascade_frontalface_alt.xml"; );
    eyes_cascade.load(  "haarcascade_eye_tree_eyeglasses.xml"; );


   capture = cvCaptureFromCAM( -1 );

  frame = cvQueryFrame( capture );
  cvtColor( frame, frame_gray, CV_BGR2GRAY );
  equalizeHist( frame_gray, frame_gray );

  face_cascade.detectMultiScale( frame_gray, faces, 1.1, 2, 0|CV_HAAR_SCALE_IMAGE, Size(30, 30) );

  for( size_t i = 0; i < faces.size(); i++ )
  {
    Point center( faces[i].x + faces[i].width*0.5, faces[i].y + faces[i].height*0.5 );
    ellipse( frame, center, Size( faces[i].width*0.5, faces[i].height*0.5), 0, 0, 360, Scalar( 255, 0, 255 ), 4, 8, 0 );

    Mat faceROI = frame_gray( faces[i] );
    std::vector<Rect> eyes;    
    eyes_cascade.detectMultiScale( faceROI, eyes, 1.1, 2, 0 |CV_HAAR_SCALE_IMAGE, Size(30, 30) );
    for( size_t j = 0; j < eyes.size(); j++ )
     {
       Point center( faces[i].x + eyes[j].x + eyes[j].width*0.5, faces[i].y + eyes[j].y + eyes[j].height*0.5 );
       int radius = cvRound( (eyes[j].width + eyes[j].height)*0.25 );
       circle( frame, center, radius, Scalar( 255, 0, 0 ), 4, 8, 0 );
     }
  }

  imshow( "window", frame );

Answer 1

Using basically your code (just changed to load a single fixed image and removed the equalizeHist call) gives me those results:

input:

output:

input:

output:

so your code basically does what it should do. Your problem seems to be in your image data or parameters regarding your image data. Post sample images please.

Just for you to be sure that I didn't change big parts of the code, I used:

using namespace cv;
using namespace std;

int main()
{

    cv::Mat input = cv::imread("../inputData/MultiLena.png");

    cv::Mat gray;
    cv::cvtColor(input,gray,CV_BGR2GRAY);


    //CvCapture* capture;
    //Mat frame;
    std::vector<Rect> faces;
    Mat frame_gray;
    Mat frame;
    CascadeClassifier face_cascade;
    CascadeClassifier eyes_cascade;

    face_cascade.load( "haarcascade_frontalface_alt.xml" );
    eyes_cascade.load(  "haarcascade_eye_tree_eyeglasses.xml" );


   //capture = cvCaptureFromCAM( -1 );

  //frame = cvQueryFrame( capture );
  //cvtColor( frame, frame_gray, CV_BGR2GRAY );
  //equalizeHist( frame_gray, frame_gray );

    frame_gray = gray;
    frame = input;


  face_cascade.detectMultiScale( frame_gray, faces, 1.1, 2, 0|CV_HAAR_SCALE_IMAGE, Size(30, 30) );

  for( size_t i = 0; i < faces.size(); i++ )
  {
    Point center( faces[i].x + faces[i].width*0.5, faces[i].y + faces[i].height*0.5 );
    ellipse( frame, center, Size( faces[i].width*0.5, faces[i].height*0.5), 0, 0, 360, Scalar( 255, 0, 255 ), 4, 8, 0 );

    Mat faceROI = frame_gray( faces[i] );
    std::vector<Rect> eyes;    
    eyes_cascade.detectMultiScale( faceROI, eyes, 1.1, 2, 0 |CV_HAAR_SCALE_IMAGE, Size(30, 30) );
    for( size_t j = 0; j < eyes.size(); j++ )
     {
       Point center( faces[i].x + eyes[j].x + eyes[j].width*0.5, faces[i].y + eyes[j].y + eyes[j].height*0.5 );
       int radius = cvRound( (eyes[j].width + eyes[j].height)*0.25 );
       circle( frame, center, radius, Scalar( 255, 0, 0 ), 4, 8, 0 );
     }
  }

  imshow( "window", frame );


  cv::imshow("input", input);
    cv::imwrite("../outputData/multiFaces.png", input);
    cv::waitKey(0);
    return 0;
}

Answer 2

maybe you could implement the following strategy:

• Detect the first face, it will generate a rect.
• Create a new Mat without that rect.
• Analyze the new mat and generate a new face detection rect.
• Now draw both rects into a video capture stream.

Keep in mind to display on screen only the background video stream with selected rects, so the system remains efficient.