Implementation of metadata generator conforming to TV-anytime specification

In this paper, we describe an implementation of indexing/presentation system conforming to TV-anytime specification, which includes metadata indexing scheme for multimedia contents, a video indexing algorithm for news and soccer videos, and MPEG-2 TS player for non-linear browsing. The overall scheme of the metadata generator is as follows. A key-frame extractor segments a video sequence into some scenes and extracts key-frame images from news and soccer videos, such as anchor frames and highlight frames. The metadata generator creates metadata documents based on the TV-anytime specifications, which include the low level information such as time duration and key-frame location extracted by key-frame extractor and the high level information such as title and synopsis written by operators.

[1]  Ki Tae Park,et al.  Anchor Frame Detection in News Video Using Anchor Object Extraction , 2005, IEICE Trans. Fundam. Electron. Commun. Comput. Sci..

[2]  Jerry D. Gibson,et al.  Handbook of Image and Video Processing , 2000 .