Human Tracking using Template Matching Methodology
Human tracking system has become a very important where security is the priority. Tracking the object in the frames from the video is a key research topic in the computer vision community. The proposed system consist of targeting human face of interest, where we investigate long-term tracking of that face in a video clips. We are using Template Matching Algorithm which is novel tracking framework (TLD) that explicitly decomposes the long-term tracking task into tracking, learning and detection. The tracker follows the object from frame to` frame. The detector localizes all appearances that have been observed so far and corrects the tracker if necessary. A P-N learning method estimates the errors by P-expert which estimates missed detections, and N-expert which estimates false alarms. The learning estimates detector’s errors and updates it to avoid these errors in the future. Experimental results shows that our approach achieves good performance on video sequences.
Keywords - Template, Learning, Detection, Bounding Box, Convolution, diffeomorphism.