News Video Indexing and Abstraction by Specific Visual Cues: MSC and News Caption

Abstract

This chapter addresses the tasks of providing the semantic structure and generating the abstraction of content in broadcast news. Based on extraction of two specific visual cues, Main Speaker Close-Up (MSC) and news caption, a hierarchy of news video index is automatically constructed for efficient access to multi-level contents. In addition, a unique MSC-based video abstraction is proposed to help satisfy the need for news preview and key-person highlighting. Experiments on news clips from MPEG-7 video content sets yield encouraging results that prove the efficiency of our video indexing and abstraction scheme.