【Svm机器学习篇】Opencv3.4.1与C++实现对多分类问题的训练与预测

#include<opencv2\opencv.hpp>
#include<iostream>
#include<string>
#include<vector>
#include<fstream>
#include<opencv2/ml/ml.hpp>

using namespace std;
using namespace cv;
using namespace ml;

//训练SVM

void svm_train(Ptr<SVM> &model, Mat &trainData, Mat &trainLabels)
{
model->setType(SVM::C_SVC); //SVM类型
model->setKernel(SVM::LINEAR); //核函数，这里使用线性核

Ptr<TrainData> tData = TrainData::create(trainData, ROW_SAMPLE, trainLabels);

   cout << "SVM: start train ..." << endl;
   model->trainAuto(tData);
   cout << "SVM: train success ..." << endl;
}

void svm_pridect(Ptr<SVM> &model, Mat test)
{
   Mat result;
   float rst = model->predict(test, result);
   for (auto i = 0; i < result.rows; i++) {
       cout << result.at<float>(i, 0);
   }
}

int main() {
   vector<Mat>image;
   //string s;
   //string str = "C:\\Users\\wangz\\Documents\\Visual Studio 2015\\Projects\\Project15\\Project15\\pic\\txt.txt";//个人路径
   string ss;
   ifstream fin("C:\\Users\\wangz\\Documents\\Visual Studio 2015\\Projects\\Project15\\Project15\\pic\\txt.txt");
   //fin.open(str);

//读取data
while (getline(fin, ss))

   {
       ss = "C:\\Users\\wangz\\Documents\\Visual Studio 2015\\Projects\\Project15\\Project15\\pic\\" + ss;
       Mat m1 = imread(ss, 0);
       resize(m1, m1, Size(64, 64));
       image.push_back(m1); //连续放入Mat容器中
   }
   Mat imagedata(200,64*64,CV_8UC1);
   for (vector<Mat>::iterator it = image.begin(); it != image.end(); it++) {
       imagedata.push_back((*it).reshape(0, 1));
   }
   imagedata.convertTo(imagedata, CV_32F);

   //读取标签
   Mat lable(200, 1, CV_8UC1);
   string sss;
   while (getline(fin, sss))

扫描二维码关注公众号，回复： 2864157 查看本文章

   {
       sss = "C:\\Users\\wangz\\Documents\\Visual Studio 2015\\Projects\\Project15\\Project15\\pic\\" + sss;

       //image.push_back(m1); //连续放入Mat容器中
   }
   Mat label(200, 64 * 64, CV_8UC1);

   for (vector<Mat>::iterator it = image.begin(); it != image.end(); it++) {
       imagedata.push_back((*it).reshape(0, 1));
   }
   imagedata.convertTo(imagedata, CV_32F);

   Ptr<SVM> model = SVM::create();
   Mat trainData, trainLabels;
   get_data(train_path, trainData, trainLabels);
   svm_train(model, trainData, trainLabels);

   Mat testData;
   transform(test_set, testData);
   svm_pridect(model, testData);

   //imshow("aa", image[0]);//显示第一张图片
   waitKey(0);
   getchar();
   return 0;

【Svm机器学习篇】Opencv3.4.1与C++实现对多分类问题的训练与预测

猜你喜欢