A comprehensive study of sports video analysis

Abstract

Commercial applications of video analysis are getting valuable with the development of digital television. People can easily record all kinds of programs and enjoy the videos in their leisure time. Among these programs, broadcasting sports videos are usually more tedious than others since they involve not only the main games, but also break time or commercials. And even main games comprise periods which are not splendid enough for the audience. Therefore, a considerable amount of research focuses on automatically annotating semantic concepts in sports videos, and providing a spellbinding way to browse videos. In this chapter, we briefly introduce related work of video analysis for different kinds of sports, and propose a generic framework for sports video annotation. We explicitly elaborate the state of the art techniques for sports videos analysis. Visual and audio information are utilized to extract mid-level features, and different models for semantic annotation are expounded with practical examples. We also expand on applications of sports video analysis from the viewpoints of the audience, professional athletes, and advertisers. © 2011 Springer-Verlag Berlin Heidelberg.