如何从OpenCv中的图像识别人

Question

I made a program that detects a face from an Image:

import org.opencv.core.Core;
import org.opencv.core.Mat;
import org.opencv.core.MatOfRect;
import org.opencv.core.Point;
import org.opencv.core.Rect;
import org.opencv.core.Scalar;
import org.opencv.highgui.Highgui;
import org.opencv.objdetect.CascadeClassifier;

class DetectFaceDemo {
public void run() {
System.out.println("\nRunning DetectFaceDemo");

CascadeClassifier faceDetector = new CascadeClassifier("C:\\Users\\HM\\Documents\\NetBeansProjects\\vision\\src\\lbpcascade_frontalface.xml");
Mat image = Highgui.imread("C:\\Users\\HM\\Downloads\\john-lennon.jpg");

MatOfRect faceDetections = new MatOfRect();
faceDetector.detectMultiScale(image, faceDetections);

System.out.println(String.format("Faces detected: %s ", faceDetections.toArray().length));

for (Rect rect : faceDetections.toArray()) {
    Core.rectangle(image, new Point(rect.x, rect.y), new Point(rect.x + rect.width, rect.y + rect.height), new Scalar(0, 255, 0), 3);
}

String filename = "detcSuccessful.png";
Highgui.imwrite(filename, image);
    }
}

public class image {
public static void main(String[] args) {
// Load the native library.
System.loadLibrary(Core.NATIVE_LIBRARY_NAME);
new DetectFaceDemo().run();
    }
}

And i got this: John lennon picture

Wanna to recognize the guy who is in the Image by the program and show his name after.

Something like: Computer: this guy is John.

Answer 1

package com.test8;

import org.opencv.core.Core;
import org.opencv.core.Mat;
import org.opencv.core.MatOfRect;
import org.opencv.core.Point;
import org.opencv.core.Rect;
import org.opencv.core.Scalar;
import org.opencv.imgcodecs.Imgcodecs;
import org.opencv.imgproc.Imgproc;
import org.opencv.objdetect.CascadeClassifier;

class FaceDetection {
    // Load the native library.
    static{ System.loadLibrary(Core.NATIVE_LIBRARY_NAME);}

    public static void main(String[] args) {

        System.out.println("\nRunning DetectFaceDemo");

        CascadeClassifier faceDetector = new CascadeClassifier("D:\\AntonKONG\\OpenCV\\opencv\\sources\\data\\haarcascades\\haarcascade_frontalface_alt.xml");
        Mat image = Imgcodecs.imread("src//data//hkid2.png");

        MatOfRect faceDetections = new MatOfRect();
        faceDetector.detectMultiScale(image, faceDetections);

        System.out.println(String.format("Faces detected: %s ", faceDetections.toArray().length));

        for (Rect rect : faceDetections.toArray()) {
            Imgproc.rectangle(image, new Point(rect.x, rect.y), new Point(rect.x + rect.width, rect.y + rect.height), new Scalar(0, 255, 0), 3);
        }

        String filename = "detcSuccessful.png";
        Imgcodecs.imwrite(filename, image);

    }
}

Answer 2

Here is the code I have previously wrote in opencv 2.4 in C++. As I see, you are using JAVA-OpenCV wrapper, so you can easily find JAVA equivalents. In the following code, is vector of images and label is vector of labels. For instace, consider your search space is limited to three persons:

0 : John Lenon 
    john1.png  // 100x90 px
    john2.png  // 100x90 px
    john3.png  // 100x90 px
1 : Robert DeNiro
    robert1.png      // 100x90 px
    robert2.png      // 100x90 px
    robert3.png      // 100x90 px
2 : AlPacino
    al1.png    // 100x90 px
    al2.png    // 100x90 px
    al3.png    // 100x90 px

0,1,2 are labels and you can see there are 3 faces correspond to each label. Read each image and store it in a Mat object. as below:

Mat j1,j2,j3; // John Lennon's faces  :: label 0
Mat r1,r2,r3; // Robert Deniro's faces :: label 1
Mat a1,a2,a3; // Al Pacino's faces :: label 2

Create your image vector and label vector as below:

images:{j1,j2,j3,r1,r2,r3,a1,a2,a3}
labels:{0,0,0,1,1,1,2,2,2}

then, pass them to the following piece of code.

cv::Ptr<cv::FaceRecognizer> model = createEigenFaceRecognizer();
    model->train(images, labels);
    int prediction;
    double confidence;
    model->predict(Frame,prediction,confidence);

Sorry for answering in C++ but I'm sure that there are similar. In addition, make sure which version of opencv you are using.