how is PCA implemented on a camera captured image?

I have successfully implemented face detection part in my Face Recognition project.Now i have a rectangular region of face in an image.Now i have to implement PCA on this detected rectangular region to extract important features.I have used examples of implementing PCA on face databases.I want to know how we can pass our detected face to function implementing PCA?Is it that we pass the rectangle frame? This is the code for my face detection.

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

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <math.h>
#include <float.h>
#include <limits.h>
#include <time.h>
#include <ctype.h>


// Create a string that contains the exact cascade name
const char* cascade_name =
    "haarcascade_frontalface_alt.xml";
/*    "haarcascade_profileface.xml";*/


// Function prototype for detecting and drawing an object from an image
void detect_and_draw( IplImage* image );

// Main function, defines the entry point for the program.
int main( int argc, char** argv )
{

    // Create a sample image
    IplImage *img = cvLoadImage("Image018.jpg");
    if(!img)
    {
        printf("could not load image");
        return -1;
    }

    // Call the function to detect and draw the face positions
    detect_and_draw(img);

    // Wait for user input before quitting the program
    cvWaitKey();

    // Release the image
    cvReleaseImage(&img);

    // Destroy the window previously created with filename: "result"
    cvDestroyWindow("result");

    // return 0 to indicate successfull execution of the program
    return 0;
}

// Function to detect and draw any faces that is present in an image
void detect_and_draw( IplImage* img )
{

    // Create memory for calculations
    static CvMemStorage* storage = 0;

    // Create a new Haar classifier
    static CvHaarClassifierCascade* cascade = 0;

    int scale = 1;

    // Create a new image based on the input image
    IplImage* temp = cvCreateImage( cvSize(img->width/scale,img->height/scale), 8, 3 );

    // Create two points to represent the face locations
    CvPoint pt1, pt2;
    int i;

    // Load the HaarClassifierCascade
    cascade = (CvHaarClassifierCascade*)cvLoad( cascade_name, 0, 0, 0 );

    // Check whether the cascade has loaded successfully. Else report and error and quit
    if( !cascade )
    {
        fprintf( stderr, "ERROR: Could not load classifier cascade\n" );
        return;
    }

    // Allocate the memory storage
    storage = cvCreateMemStorage(0);

    // Create a new named window with title: result
    cvNamedWindow( "result", 1 );

    // Clear the memory storage which was used before
    cvClearMemStorage( storage );

    // Find whether the cascade is loaded, to find the faces. If yes, then:
    if( cascade )
    {

        // There can be more than one face in an image. So create a growable sequence of faces.
        // Detect the objects and store them in the sequence
        CvSeq* faces = cvHaarDetectObjects( img, cascade, storage,
                                            1.1, 2, CV_HAAR_DO_CANNY_PRUNING,
                                            cvSize(40, 40) );

        // Loop the number of faces found.
        for( i = 0; i < (faces ? faces->total : 0); i++ )
        {
           // Create a new rectangle for drawing the face
            CvRect* r = (CvRect*)cvGetSeqElem( faces, i );

            // Find the dimensions of the face,and scale it if necessary
            pt1.x = r->x*scale;
            pt2.x = (r->x+r->width)*scale;
            pt1.y = r->y*scale;
            pt2.y = (r->y+r->height)*scale;

            // Draw the rectangle in the input image
            cvRectangle( img, pt1, pt2, CV_RGB(255,0,0), 3, 8, 0 );
        }
    }

    // Show the image in the window named "result"
    cvShowImage( "result", img );

    // Release the temp image created.
    cvReleaseImage( &temp );
}

Edit:

Just to notify anyone visiting this site. I have written some sample code to perform face recognition in videos using my libfacerec library:

https://github.com/bytefish/libfacerec/blob/master/samples/facerec_video.cpp

Original post:

I assume your problem is the following. You've used the Cascade Classifier cv::CascadeClassifier coming with OpenCV to detect and extract faces from images. Now you want to perform a face recognition on the images.

You want to use the Eigenfaces for face recognition. So the first thing you have to do is to learn the Eigenfaces from the images you've gathered. I rewrote the Eigenfaces class for you to make it simpler. To learn the eigenfaces simply pass a vector with your face images and the corresponding labels (the subject) either to Eigenfaces::Eigenfaces or Eigenfaces::compute. Make sure all your images have the same size, you can use cv::resize to ensure this.

Once you have computed the Eigenfaces, you can get predictions from your model. Simply call Eigenfaces::predict on a computed model. The main.cpp shows you how to use the class and its methods (for prediction, projection, reconstruction of images), here's how to get a prediction for an image.

Now I see where your problem is. You are using the old OpenCV C API. That makes it's hard to interface with the new OpenCV2 C++ API my code is written in. Not to be offending, but if you want to interface with my code you better use the OpenCV2 C++ API. I can't give a guide on learning C++ and the OpenCV2 API here, there's a lot of documentation coming with OpenCV. A good start is the OpenCV C++ Cheat Sheet (also available at http://opencv.willowgarage.com/) or the OpenCV Reference Manual.

For recognizing images from the Cascade Detector, I repeat: First learn the Eigenfaces model with the persons you want to recognize, it's shown in the example coming with my code. Then you need to get the Region Of Interest (ROI), that's the face, the Rectangle the Cascade Detector outputs. Finally you can get a prediction for the ROI from the Eigenfaces model (you have computed it above), it's shown in the example coming with my code. You probably have to convert your image to grayscale, but that's all. That's how it's done.