Opencv3.x中SVM+HOG训练测试代码

其实训练代码还是比较简单的，设置完HOG和SVM的参数就可以开始训练了。需要训练时就不用注释#define TRAIN，若想运行之前已经训练好的模型则需将其注释掉，并修改模型名。

#include <opencv2/core/core.hpp>
#include <opencv2/highgui/highgui.hpp>
#include <opencv2/objdetect/objdetect.hpp>
#include "opencv2/imgproc/imgproc.hpp"
#include "opencv2/highgui/highgui.hpp"
#include <fstream>
#include <iostream>
#include <opencv2/ml/ml.hpp>
#define TRAIN
using namespace std;
using namespace cv::ml;
using namespace cv;

Size imageSize = Size(96, 128);

void coumputeHog(const Mat& src, vector<float> &descriptors){
	//winsize  blocksize  blockStride  cellsize  bins
	HOGDescriptor myHog = HOGDescriptor(imageSize, Size(96, 64), Size(96, 32), Size(48, 32), 4);
	myHog.compute(src.clone(),descriptors,Size(1,1),Size(0,0));
}
 
int main(int argc, char** argv){
	ifstream inLabels("myImageLabels1.txt"), inImages("myImageList1.txt"),inTestimage("myImagetest1.txt");
	string imageName;
	signed imageLabel;
	vector<Mat> vecImages;
	vector<int> vecLabels;
	vector<float> vecDescriptors;

#ifdef TRAIN
	Ptr<SVM> mySVM = SVM::create();

	mySVM->setType(SVM::C_SVC);
	mySVM->setKernel(SVM::LINEAR);
	mySVM->setTermCriteria(TermCriteria(TermCriteria::MAX_ITER + TermCriteria::EPS,1500, 0.01));
	mySVM->setC(450);

	
	while ((inImages>>imageName)&&(inLabels>>imageLabel))
	{
		Mat src = imread(imageName, 0);
		resize(src, src, imageSize);
		vecImages.push_back(src);
		vecLabels.push_back(imageLabel);
	}
	inLabels.close();
	inImages.close();
 
	Mat dataDescriptors;
	Mat dataResponse = (Mat)vecLabels;
	for (size_t i = 0; i < vecImages.size(); i++)
	{
		Mat src = vecImages[i];
		Mat tempRow;
		coumputeHog(src, vecDescriptors);
		cout<<vecDescriptors.size()<<endl;
		if (i == 0)
		{
			dataDescriptors = Mat::zeros(vecImages.size(), vecDescriptors.size(), CV_32FC1);
		}
		tempRow = ((Mat)vecDescriptors).t();
		tempRow.row(0).copyTo(dataDescriptors.row(i));

	}
	Ptr<TrainData> tData = TrainData::create(dataDescriptors, ROW_SAMPLE, dataResponse);
	mySVM->train(tData);
	string svmName = to_string(1) + "_hogsvm.xml";
	mySVM->save(svmName.c_str());	
#endif

#ifndef TRAIN
	Ptr<ml::SVM> mySVM = Algorithm::load<ml::SVM>("0hogsvm.xml");
#endif  
	string testPath;
	while (inTestimage >> testPath){
		Mat test = imread(testPath, 0);
		resize(test, test, imageSize);
		vector<float> imageDescriptor;
		coumputeHog(test, imageDescriptor);
		Mat testDescriptor = Mat::zeros(1, imageDescriptor.size(), CV_32FC1);
		for (size_t i = 0; i < imageDescriptor.size(); i++){
			testDescriptor.at<float>(0, i) = imageDescriptor[i];
		}		
		float  label = mySVM->predict(testDescriptor);
		cout << label << endl;
		imshow("test image", test);
		//waitKey(0);	
	}
 
	inTestimage.close();
	delete mySVM;
	return 0;
}

myImageLabels1.txt  训练图片对应的标签

myImageList1.txt  训练图片对应的路径

myImagetest1.txt  测试图片对应的路径

  如果路径和标签过多的话，txt里面的内容那就得靠代码帮你自动化完成了....