OpenCV SVM

#include <opencv2/core/core.hpp>  
#include <opencv2/highgui/highgui.hpp>  
#include <opencv2/ml/ml.hpp>  

using namespace cv;

int main()
{
    // Data for visual representation  
    int width = 512, height = 512;
    Mat image = Mat::zeros(height, width, CV_8UC3);

    // Set up training data  
    float labels[5] = { 1.0, -1.0, -1.0, -1.0, 1.0 };
    Mat labelsMat(5, 1, CV_32FC1, labels);


    float trainingData[5][2] = { { 501, 10 }, { 255, 10 }, { 501, 255 }, { 10, 501 }, { 501, 128 } };
    Mat trainingDataMat(5, 2, CV_32FC1, trainingData);

    //设置支持向量机的参数  
    CvSVMParams params;
    params.svm_type = CvSVM::C_SVC;//SVM类型：使用C支持向量机  
    params.kernel_type = CvSVM::LINEAR;//核函数类型：线性  
    params.term_crit = cvTermCriteria(CV_TERMCRIT_ITER, 100, 1e-6);//终止准则函数：当迭代次数达到最大值时终止  

    //训练SVM  
    //建立一个SVM类的实例  
    CvSVM SVM;
    //训练模型，参数为：输入数据、响应、XX、XX、参数（前面设置过）  
    SVM.train(trainingDataMat, labelsMat, Mat(), Mat(), params);

    Vec3b green(0, 255, 0), blue(255, 0, 0);
    //显示判决域  
    for (int i = 0; i < image.rows; ++i)
    for (int j = 0; j < image.cols; ++j)
    {
        Mat sampleMat = (Mat_<float>(1, 2) << i, j);
        //predict是用来预测的，参数为：样本、返回值类型（如果值为ture而且是一个2类问题则返回判决函数值，否则返回类标签）、  
        float response = SVM.predict(sampleMat);

        if (response == 1)
            image.at<Vec3b>(j, i) = green;
        else if (response == -1)
            image.at<Vec3b>(j, i) = blue;
    }

    //画出训练数据  
    int thickness = -1;
    int lineType = 8;
    circle(image, Point(501, 10), 5, Scalar(0, 0, 0), thickness, lineType);//画圆  
    circle(image, Point(255, 10), 5, Scalar(255, 255, 255), thickness, lineType);
    circle(image, Point(501, 255), 5, Scalar(255, 255, 255), thickness, lineType);
    circle(image, Point(10, 501), 5, Scalar(255, 255, 255), thickness, lineType);
    circle(image, Point(501, 128), 5, Scalar(0, 0, 0), thickness, lineType);

    //显示支持向量  
    thickness = 2;
    lineType = 8;
    //获取支持向量的个数  
    int c = SVM.get_support_vector_count();

    for (int i = 0; i < c; ++i)
    {
        //获取第i个支持向量  
        const float* v = SVM.get_support_vector(i);
        //支持向量用到的样本点，用灰色进行标注  
        circle(image, Point((int)v[0], (int)v[1]), 6, Scalar(128, 128, 128), thickness, lineType);
    }

    imwrite("result.png", image);        // save the image   

    imshow("SVM Simple Example", image); // show it to the user  
    waitKey(0);

}