Fast face clustering using similar shots for face navigation of TV programs

In this paper, we propose a new approach for clustering faces of characters in a recorded television title. The clustering results are used to catalog video clips based on subjects' faces for quick scene access. The main goal is to obtain a reasonable result for cataloging in a short time as soon as the recording phase is over. To enable high-speed processing, similarities of shots where the characters appear are used to estimate corresponding faces instead of calculating distance between each face feature. Two kinds of experiments are conducted to evaluate the method. With the fastest method, processing time is less than 1 second per hour of video clips, which is 1000 times faster than in the case of using rigorous face verification with facial feature extraction. The difference between these methods when used to catalog video clips was within allowable error.