How to get more accuracy for Face Recognition with images?
Hi, I am implementing Face Recognition project with Eigenfaces and Fisherfaces algorithms in c++. We are use here 5-8 patch images of same single person for matching face in group photo. But accuracy of matching single person faces in group photo with patch images is very less. I get only 20-25 % accuracy. In 100 group photos algorithm recognize person is present in only 22-28 photos accurately. How to improve or get more accuracy for Face Recognition? Which framework or algorithm are more useful for Face Recognition with more accuracy?
I guess you should start with a search on this forum, where this question has been raised for a thousand times. Furthermore, in order to increase efficiency, you want a lot more samples per person, like at least 50.