Video Cataloguing and Browsing

Tandianus, J.E., Chandra, A. and Jin, J.S.

Videos contain rich information. Until recently, information within a video had been largely left under-utilized, with fast forward/rewind as the most popular method of accessing video content. This is no longer sufficient however, as more and more users demanded the flexibility and ability to access video content selectively. Unlike textual information however, video bits do not convey same level of meaning as text do. Therefore, we have to use metadata to describe the structured information within videos. One project, ViMeta-Vu or Video Metadata Viewer, was undertaken to provide users with such content-based video access. It represented video using fixed three-level hierarchical structure. Users can access and retrieve specific video information through browsing of the metadata This paper will discuss various ways of improving the existing system. It proposes a more generic hierarchical structure using XML to represent the metadata, which will facilitate additional query methods. This paper also proposes additional query methods utilizing visual features of key frames in the video. The queries implemented include structural browsing, query by feature, query by selection and query by keyword.

Cite as: Tandianus, J.E., Chandra, A. and Jin, J.S. (2002). Video Cataloguing and Browsing. In Proc. Selected papers from 2001 Pan-Sydney Area Workshop on Visual Information Processing (VIP2001), Sydney, Australia. CRPIT, 11. Feng, D. D., Jin, J., Eades, P. and Yan, H., Eds. ACS. 39-45.

(from crpit.com) (local if available)