This is my code.... It works well but, the classifier is not predicting right.. it shows the same result each time for any kind of evaluation data.. I am not able to figure out where the problem lies. Its very urgent. I'll be grateful
include "stdafx.h"
include <vector>
include <boost filesystem.hpp="">
include <opencv2 opencv.hpp="">
include<stdio.h>
include<conio.h>
using namespace std; using namespace boost::filesystem; using namespace cv;
//location of the training data
define TRAINING_DATA_DIR "data/train/"
//location of the evaluation data
define EVAL_DATA_DIR "data/eval/"
//See article on BoW model for details
Ptr<featuredetector> detector = FeatureDetector::create("SURF"); Ptr<descriptorextractor> extractor = DescriptorExtractor::create("SURF"); Ptr<descriptormatcher> matcher = DescriptorMatcher::create("FlannBased"); //See article on BoW model for details int dictionarySize = 1000; TermCriteria tc(CV_TERMCRIT_ITER+CV_TERMCRIT_EPS, 10, 0.001); int retries = 1; int flags = KMEANS_PP_CENTERS;
//See article on BoW model for details BOWKMeansTrainer bowTrainer(dictionarySize, tc, retries, flags); //See article on BoW model for details BOWImgDescriptorExtractor bowDE(extractor, matcher); /* CvSVMParams params; params.svm_type = CvSVM::C_SVC; params.kernel_type = CvSVM::LINEAR; params.term_crit = cvTermCriteria(CV_TERMCRIT_ITER, 100, 1e-6); / /* * \brief Recursively traverses a folder hierarchy. Extracts features from the training images and adds them to the bowTrainer. */
bool readVocabulary( const string& filename, Mat& vocabulary ) { cout << "Reading vocabulary..."; FileStorage fs( filename, FileStorage::READ ); if( fs.isOpened() ) { fs["vocabulary"] >> vocabulary; cout << "done" << endl; return true; } return false; }
bool writeVocabulary( const string& filename, const Mat& vocabulary ) { cout << "Saving vocabulary..." << endl; FileStorage fs( filename, FileStorage::WRITE ); if( fs.isOpened() ) { fs << "vocabulary" << vocabulary; return true; } return false; }
void extractTrainingVocabulary(const path& basepath) { for (directory_iterator iter = directory_iterator(basepath); iter != directory_iterator(); iter++) { directory_entry entry = *iter;
if (is_directory(entry.path())) {
cout << "Processing directory " << entry.path().string() << endl;
extractTrainingVocabulary(entry.path());
} else {
path entryPath = entry.path();
if (entryPath.extension() == ".jpg") {
cout << "Processing file " << entryPath.string() << endl;
Mat img = imread(entryPath.string(),0);
if (!img.empty()) {
vector<KeyPoint> keypoints;
detector->detect(img, keypoints);
if (keypoints.empty()) {
cerr << "Warning: Could not find key points in image: "
<< entryPath.string() << endl;
} else {
Mat features;
extractor->compute(img, keypoints, features);
bowTrainer.add(features);
}
} else {
cerr << "Warning: Could not read image: "
<< entryPath.string() << endl;
}
}
}
}
}
/** * \brief Recursively traverses a folder hierarchy. Creates a BoW descriptor for each image encountered. */ void extractBOWDescriptor(const path& basepath, Mat& descriptors, Mat& labels) { for (directory_iterator iter = directory_iterator(basepath); iter != directory_iterator(); iter++) { directory_entry entry = *iter; if (is_directory(entry.path())) { cout << "Processing directory " << entry.path().string() << endl; extractBOWDescriptor(entry.path(), descriptors, labels); } else { path entryPath = entry.path(); if (entryPath.extension() == ".jpg") { cout << "Processing file " << entryPath.string() << endl; Mat img = imread(entryPath.string(),0); if (!img.empty()) { vector<keypoint> keypoints; detector->detect(img, keypoints); if (keypoints.empty()) { cerr << "Warning: Could not find key points in image: " << entryPath.string() << endl; } else { Mat histogram; bowDE.compute(img, keypoints, histogram); descriptors.push_back(histogram); float label=atof(entryPath.filename().string().c_str()); labels.push_back(label); } } else { cerr << "Warning: Could not read image: " << entryPath.string() << endl; } } } }
}
int main(int argc, char ** argv) {
cout<<"Creating dictionary..."<<endl;
Mat dictionary;
if( !readVocabulary( "k.txt", dictionary) )
{
extractTrainingVocabulary(path(TRAINING_DATA_DIR));
vector<Mat> descriptors = bowTrainer.getDescriptors();
int count=0;
for(vector<Mat>::iterator iter=descriptors.begin();iter!=descriptors.end();iter++)
{
count+=iter->rows;
}
cout<<"Clustering "<<count<<" features"<<endl;
dictionary = bowTrainer.cluster();
if( !writeVocabulary("k.txt", dictionary) )
{
cout << "Error: file k.txt" << " can not be opened to write" << endl;
exit(-1);
}
}
else
bowDE.setVocabulary(dictionary);
cout<<"Processing training data..."<<endl;
Mat trainingData(0, dictionarySize, CV_32FC1);
Mat labels(0, 1, CV_32FC1);
extractBOWDescriptor(path(TRAINING_DATA_DIR), trainingData, labels);
if( !writeVocabulary("Histogram.txt", trainingData) )
{
cout << "Error: file Histogram.txt" << " can not be opened to write" << endl;
exit(-1);
}
CvNormalBayesClassifier classifier;
classifier=new CvNormalBayesClassifier();
Mat temp;
cout<<"Training classifier..."<<endl;
if( !readVocabulary( "BayesDictalology.xml", temp) ){
classifier.train(trainingData, labels);
classifier.save("BayesDictalology.xml","BayesDictalology");
}
else
classifier.load("BayesDictalology.xml","BayesDictalology");
//CvSVM SVM;
//SVM.train(trainingData,labels);
cout<<"Processing evaluation data..."<<endl;
Mat evalData(0, dictionarySize, CV_32FC1);
Mat groundTruth(0, 1, CV_32FC1);
extractBOWDescriptor(path(EVAL_DATA_DIR), evalData, groundTruth);
cout<<"Evaluating classifier..."<<endl;
Mat results;
//cout<<SVM.predict(evalData);
classifier.predict(evalData, &results);
int rows=results.rows;
int columns=results.cols;
for(int i=0;i<rows;i++)
{
for(int j=0;j<columns;j++)
cout<<" "<<results.at<float>(i,j);
cout<<endl;
}
_getche();
double errorRate = (double) countNonZero(groundTruth - results) / evalData.rows;
;
cout << "Error rate: " << errorRate << endl;
_getche();
}
