implementation of bayes and svm classifiers in opencv3

Question

I'd like to implement Bayes and SVM classifiers in OpenCV 3.3. I've written following code for SVM:

int main()
{
// Load data

FileStorage fs("newStorageFile.yml", FileStorage::READ);

// Read data

Mat test_data, test_labels, enrol_data, enrol_labels;
Mat train_labels = Mat::zeros(650, 1, CV_32S);
Mat train_data = Mat::zeros(650, 600, CV_32S);

fs["train_data"] >> train_data;
fs["train_labels"] >> train_labels;
fs["test_data"] >> test_data;
fs["test_labels"] >> test_labels;
fs["enrol_data"] >> enrol_data;
fs["enrol_labels"] >> enrol_labels;


Ptr<ml::SVM> SVM_Model = ml::SVM::create();
SVM_Model->setType(ml::SVM::C_SVC);
SVM_Model->setKernel(ml::SVM::RBF);

Ptr<ml::TrainData> trainingData = ml::TrainData::create(train_data, ml::SampleTypes::ROW_SAMPLE, train_labels);
SVM_Model->trainAuto(trainingData);

return 0;
}

But I've got following exception error on SVM_Model->trainAuto(trainingData).

Unhandled exception at 0x755B5608 in SVMimplemantation.exe: Microsoft C++ exception: cv::Exception at memory location 0x00D8DF58.

And also about Bayes classifier, I've written following code:

int main()
{
// Load data

FileStorage fs("newStorageFile.yml", FileStorage::READ);

// Read data

Mat test_data, test_labels, enrol_data, enrol_labels;
Mat train_labels = Mat::zeros(650, 1, CV_32F);
Mat train_data = Mat::zeros(650, 600, CV_32F);

fs["train_data"] >> train_data;
fs["train_labels"] >> train_labels;
fs["test_data"] >> test_data;
fs["test_labels"] >> test_labels;
fs["enrol_data"] >> enrol_data;
fs["enrol_labels"] >> enrol_labels;

Ptr<ml::NormalBayesClassifier> bayes = ml::NormalBayesClassifier::create();
Ptr<ml::TrainData> trainData = ml::TrainData::create(train_data, ml::SampleTypes::ROW_SAMPLE, train_labels);
bayes->train(trainData);

Mat output, outputProb;

bayes->predictProb(test_data, output, outputProb);

return 0;
}

About this case, I've got following Exception on bayes->train(trainData) too.

Unhandled exception at 0x755B5608 in BayesImplementation.exe: Microsoft C++ exception: cv::Exception at memory location 0x00A5D79C.

To be able to compile the project, Ive uploaded my dataset here. What's the problem and how to fix it?

Have you checked all your pointers? All the create functions that return Ptr<> will return false on failure. Check after each create with something like if (!trainData) throw("Bad pointer"); Edit: That is, they won't return false, but the Ptr<> class has a bool conversion so you can check if it's valid. — Zebrafish
– Zebrafish, Commented Mar 8, 2018 at 7:59
@Zebrafish thanks, I've checked it. but the problem is somewhere else! — DavRay
– DavRay, Commented Mar 8, 2018 at 8:37

Saeid · Accepted Answer · 2018-03-08 11:38:28Z

1

Problem is the type of train_label. Try with following:

Mat train_labels32S = Mat::zeros(train_labels.rows, 1, CV_32S);

for (int i = 0; i < train_labels.rows; i++)
    train_labels32S.at<int>(i, 0) = train_labels.at<int>(i, 0);
//some code
    Ptr<ml::TrainData> trainingData = ml::TrainData::create(train_data, ml::SampleTypes::ROW_SAMPLE, train_labels32S);

answered Mar 8, 2018 at 11:38

Saeid

5781 gold badge5 silver badges21 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

implementation of bayes and svm classifiers in opencv3

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related