Repository | Book | Chapter

The growth of digital broadcasting has lead to the emergence and wide spread distribution of TV terminals that are equipped with set-top boxes (STB) and personal video recorders (PVR). With increasing number of broadcasting channels and broadcasting services becoming more personalized, today's TV terminal requires more complex structures and functions such as picture-in-picture, time-shift play, channel recording, etc. The increasing number of channels also complicates the efforts of TV viewers in finding their favorite broadcasts quickly and efficiently. In addition, obtaining meaningful scenes from live broadcasts becomes more difficult as the metadata describing the broadcasts is not available [1]. In fact, most live broadcasts cannot afford to provide related metadata as it must be prepared before the broadcast is aired.Until now, many scene detection and content-indexing techniques have been applied to video archiving systems for video summarization, video segmentation, content management, and metadata authoring of broadcasts [2–7]. Some of them have been applied in STB or PVR after recording entire broadcasts [8–10]. Current broadcasting systems provide simple program guiding services with electronic program guides, but do not provide meaningful scene searching for live broadcasts [11, 12]. In order to provide a user-customized watching environment in digital broadcasting, meaningful scene search in real-time is required in the TV terminal. N. Dimitrova et al. [13], studied video analysis algorithms and architecture for abstracted video representation in consumer domain applications and developed a commercial tool for skipping through the detection of black frames and changes in activity [14]. Our work, however, focuses on establishing a system that enables the indexing and analyzing of live broadcasts at the content-level.

DOI: 10.1007/978-0-387-89024-1_18

Full citation: