Automatic classification of tennis video for content-based retrieval

Abstract

Automatic classification and retrieval of video information based on content is a very challenging research area. It is generally accepted that content-based retrieval for generic video is extremely difficult, limiting the analyses to specific domains can help overcome many hurdles in automatic generation of high-level annotations as relevant to those specific domains. The paper presents a case study of tennis video. We build a real time online automatic classification system for tennis video to facilitate content-based retrieval. We develop court-line detection algorithms to automatically extract tennis court. A player tracking algorithm which does not use template technique is designed to track tennis players. By analysis of the relative position of the tennis players with respect to the court lines and the net, our system can map high-level events like baseline-rallies, server-and-volleying, net-games, passing-shots, etc. Results on real tennis video data are presented to demonstrate the validity of the approach.

Date of Award	2000
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology