The VISION Digital Video Library System
Project Award Date: 0000-00-00
This project has developed a prototypical digital video libary system called VISION (Video Indexing for Searching Over Networks). VISION can digitize, compress, store, index, search, and retrieve video, audio, and closed-caption information. We have developed a client and server which can provide video information over the Internet or the evolving National Information Infrastructure.
To be an effective library, users need to be able to find the video segments they want. This project is conducting ground-breaking research into automatic content-based segmentation and indexing of video that will significantly improve the users' ability to access specific segments of interest with videos.
Videos, soundtracks and closed-captions are digitized, and the video and audio signals are used together to provide scene-based segementation rather than pixel-based segmentation. Information form the closed-captions are used to automatically index the segments based on their contents. Together, this allows users to quickly search indices for multiple videos to locate segments of interest, and to view and manipulate these segments on their remote computer. Current work is investigating the use of soundtrack and transcripts for further indexing information.
For More Information: http://www.ittc.ku.edu/~sgauch/DVLS.html
Primary Sponsor(s): Information and Telecommunication Technology Center (ITTC), DesignLab, University of Kansas Research Development Fund