%0 Conference Proceedings %F dvmmPub156 %A Chang, Shih-Fu %A Smith, John R. %A Jianhao Meng, Horace %T Efficient Techniques for Feature-Based Image/Video Access and Manipulation %B 33rd Annual Clinic on Library Applications of Data Processing - Digital Image Access and Retrieval (Invited Paper) %C University of Illinois at Urbana-Champaign, IL %X In this paper we describe two example prototype systems of CBVQ, VisualSEEk and CVEPS. VisualSEEk is a fully automated content-based image query system which allows users to search images by localized colors and their spatial layout. It includes a Java-based Web interface for interactive visual content specification, integrated visual/textual search, and performance evaluation. It's the first CBVQ system which supports local content specification and integrated query using visual feature and spatial layout. CVEPS (Compressed Video Editing and Parsing System) is a software prototype of video indexing and manipulation system which support automatic video segment decomposition, video indexing based on key frames or objects, and compressed video editing. It's the first software-based system supporting both indexing and manipulation of compressed video. The compressed domain approach provide great benefits in reducing computational complexity, storage space requirement, and improving picture quality. Implementation of a simple example editing function, random cut and paste, can be improved in speed by about 100 times by using the compressed-domain approach %U http://www.ee.columbia.edu/dvmm/publications/96/chang96b.pdf %8 March %D 1996