Research and develope of badminton sports video retrieval system

Content-based sports video retrieval becomes an important branch of multimedia retrieval which is free from traditional manual text labeling. A content-based retrieval system of Badminton sports video based on visual feature is researched and developed in this paper. A Badminton video is first automatically segmented into many sub-clips based on visual features, and then similar video clips can be obtained using sliding sub-clip window algorithm and matching function, among these sub-clips, similar video clips are ranked by a sorting model. The experimental results showed that our system could effectively and efficiently retrieve similar Badminton sports video corresponding with a query clip.