Fastest way to track multiple faces?

asked 2016-10-17 01:40:09 -0500

Hi, everyone! I am doing a project that the program can detect faces in the video. And you know in detection, it could not always find a face during the whole process, so i want to use tracking to help me detect the face. The problem is even i use the KCF algorithm to track multiple faces, the time would be more and more depending on the number of the faces. Therefore, i'd like to know if there any other way to track multiple faces in a short time?

edit retag flag offensive close merge delete


I have never had any performance issues with LBP cascades in real-time. Is that one too slow for you?

benT gravatar imagebenT ( 2016-10-18 10:53:04 -0500 )edit