CODEWITHSUNDEEP
pythonopencvsvm
user2233244
user2233244

Reputation: 427

Classification using SVM from opencv

I have problem with classification using SVM. Let's say that I have 10 classes, digts from 0 to 9. I can train SVM to recognize theese classes, but sometimes I get image which is not digt, but SVM still tries to categorize this image. Is there a way to set threshold for SVM on the output maybe (as I can set it for Neural Networks) to reject bad images? May I ask for code sample (in C++ or Python with opencv)? Thanks in advance.

Upvotes: 0

Views: 691

Answers (1)

Яois
Яois

Reputation: 3858

As a simple approach, you can train an additional classifier to determine if your feature is a digit or not. Use non-digit images as positive examples and the other classes' positives (i.e. images of digits 0-9) as the negative samples of this classifier. You'll need a huge amount of non-digit images to make it work, and also it's recommendable to use strategies as the selection of hard negatives: negative samples classified as "false positives" after the first training stage, which are used to re-train the classifier.

Hope that it helps!

Upvotes: 1

Related Questions