gistfile1.txt

#include "opencv2/opencv.hpp"


#include <iostream>
#include <fstream>
#include <ctype.h>

using namespace cv;
using namespace std;


int main(int argc, char* argv[])
{
    double varX=2,varY=4;
    float scale=10;
    int nbData=400;
    Mat h=(Mat_<float>(2,3)<<scale,0,7*varX*scale,0,-scale, 7 * varY*scale);
    Mat groups;
    Mat samples(nbData, 2, CV_32F);
    RNG r;
    Mat img(int(2*h.at<float>(1,2)),int(2*h.at<float>(0, 2) ),CV_8UC3,Scalar::all(0));
    Rect rImg(0,0,img.cols,img.rows);
    int nbClasse=1;
    for (int i = 0; i < nbData; ++i)
    {
        int ind=0;
        if (i < nbData )
        {
            samples.at<float>(i, 0) = r.gaussian(varX);
            samples.at<float>(i, 1) = r.gaussian(varY);
            ind=0;

        }
        else
        {
            samples.at<float>(i, 0) = r.gaussian(varX/2)+5;
            samples.at<float>(i, 1) = r.gaussian(varY/2)+5;
            ind=1;
            nbClasse=2;

        }
        groups.push_back(ind);

        Mat ph(3,1,CV_32F,Scalar(1));
        ph(Range(0,2),Range(0,1))=samples.row(i).t();
        Mat p=h*ph;
        if (rImg.contains(Point(p)))
            if (ind)
                circle(img, Point(p.at<float>(0, 0), p.at<float>(1, 0)), 2, Scalar(0, 0, 255), 2);
            else
                circle(img, Point(p.at<float>(0, 0), p.at<float>(1, 0)), 2, Scalar(255, 0, 255), 2);

    }
    cout << samples << endl;
    cout << groups << endl;
    Ptr<ml::SVM> classifierSVM = ml::SVM::create();
    if (nbClasse==2)
        classifierSVM->setType(ml::SVM::C_SVC);
    else
        classifierSVM->setType(ml::SVM::ONE_CLASS);
    classifierSVM->setKernel(ml::SVM::LINEAR);
    classifierSVM->setDegree(3);
    classifierSVM->setGamma(1);
    classifierSVM->setCoef0(0);
    classifierSVM->setC(1);
    classifierSVM->setNu(0.1);
    classifierSVM->setP(0);
    classifierSVM->setTermCriteria(cvTermCriteria(CV_TERMCRIT_ITER + CV_TERMCRIT_EPS,500, FLT_EPSILON));

    classifierSVM->train(samples, ml::ROW_SAMPLE, groups);
    classifierSVM->save("trainingData.yml");

    for (int i = 0; i < 20; ++i) {
        Mat test(1, 2, CV_32F);
        if (i%2==0)
        {
            test.at<float>(0, 0) = float(i-10);
            test.at<float>(0, 1) = float(i-10);

        }
        else
            test = samples.row(i);
        float result = classifierSVM->predict(test);
        cout << test << ": class " << result << endl;
        Mat ph(3, 1, CV_32F, Scalar(1));
        ph(Range(0, 2), Range(0, 1)) = test.t();
        Mat p = h*ph;
        if (rImg.contains(Point(p)))
        {
            if (result==0)
                rectangle(img, Rect(p.at<float>(0, 0) - 1, p.at<float>(1, 0) - 1, 3, 3), Scalar(0, 255, 0), 2);
            else
                rectangle(img, Rect(p.at<float>(0, 0) - 1, p.at<float>(1,0) - 1, 3, 3), Scalar(255, 255, 255), 2);

        }
    }
    imshow("SVM",img);
    waitKey(0);
    return 0;
}