Abstract: The main objective of this paper is to propose a content based video retrieval model for effective and efficient retrieval of video. The proposed model integrates the temporal and spatial features of video using motion activity descriptors and dominant color descriptors. The motion intensity and color descriptors that are MPEG-7 standardized are combined for the indexing and retrieval of video. The motion descriptor is computed from P frames of video and the dominant colors are also computed for that frame as a descriptor. Integrating dominant color descriptor with motion descriptors achieves a better retrieval of video.
Seethalakshmi, P , V. Ramachandran and R. Pandian , 2004. An Enhanced Model for Content Based Video Retrieval . Asian Journal of Information Technology, 3: 1284-1288.