Weighted block matching-based anchor shot detection with dynamic background

Abstract

Anchor shot detection is fundamental to the structure analysis of news program, such as story segmentation and commercial block location. Previous work mainly tackled the problem of picture in frame on half screen, while there are also some news programs which have a global dynamic studio background. In this paper, a new block weighing based anchor shot detection algorithm is proposed. Firstly, the key frame of each shot is extracted and divided into 64 blocks equally. Then a weighted template is built automatically after clustering analysis of different types of shots, which indicates the degrees of variation on the different blocks in the anchor shot background. And lastly, all anchor shots are detected by a weighted template matching scheme. The detection results have shown that our method is effective, especially suitable for the news program which has a global dynamic background. © 2009 Springer Berlin Heidelberg.