Compressed domain video retrieval using object and global motion descriptors

Video content description has become an important task with the standardization effort of MPEG-7, which aims at easy and efficient access to visual information. In this paper we propose a system to extract object-based and global features from compressed MPEG video using the motion vector information for video retrieval. The reliability of the motion information is enhanced by a motion accumulation process. The global features like motion activity and camera motion parameters are extracted from the above enhanced motion information. The object features such as speed, area and trajectory are then obtained after the proposed object segmentation. The number of objects in a given video shot is determined by the proposed K-means clustering procedure. The object segmentation is done by applying EM algorithm.