Digital Video Analysis HONGJIANG Institute and Recognition for Corltent-Based Access ZHANG Science, and QI TIAN National University of Singapore, ( tian@iss.nus.sg )< hjzhang@hpl.hp.tom) of Systems Digital video media analysis and recognition (DVMAR) has become an active research topic in multimedia systems and computer vision areas. This is because progress of computer and communication technologies has created strong demand for many applications of digital video in a wide variety of areas, many of which require managing video in a database (mostly distributed) or information system environment; on the other hand, indexing and retrieval schemes for traditional databases and video manipulation tools in current TV systems cannot manage video in an effective, interactive, and content-based manner [Gudivada et al. 1995]. The goal of DVMAR is to develop algorithms, tools, and systems to extract and analyze basic elements, features, and structures (syntactic and semantic) of video so as to make content-based access and transmission of video data feasible and more effective [Gudivada et al. 1995; Zhang et al. 1995]. Let us first review how a video program is organized and how DVMAR is performed at various video component levels. In general, video media is considered to have the following basic structural
/lp/association-for-computing-machinery/digital-video-analysis-and-recognition-for-content-based-access-GAX9E9cJpt