An Overview of Video Shot Clustering and Summarization Techniques for Mobile Applications

The problem of content characterization of video programmes is of great interest because video appeals to large audiences and its efficient distribution over various networks should contribute to widespread usage of multimedia services. In this paper we analyze several techniques proposed in literature for content characterization of video programmes, including movies and sports, that could be helpful for mobile media consumption. In particular we focus our analysis on shot clustering methods and effective video summarization techniques since, in the current video analysis scenario, they facilitate the access to the content and help in quick understanding of the associated semantics. First we consider the shot clustering techniques based on low-level features, using visual, audio and motion information, even combined in a multi-modal fashion. Then we concentrate on summarization techniques, such as static storyboards, dynamic video skimming and the extraction of sport highlights. Discussed summarization methods can be employed in the development of tools that would be greatly useful to most mobile users: in fact these algorithms automatically shorten the original video while preserving most events by highlighting only the important content. The effectiveness of each approach has been analyzed, showing that it mainly depends on the kind of video programme it relates to, and the type of summary or highlights we are focusing on.