Video retrieval based on scene change detection in compressed streams

This paper presents a new key frame extraction algorithm as well as a novel video-indexing scheme for fast content-based browsing and retrieval in video databases. We first extract key frames by matching the AHIM (accumulated histogram intersection measure) of the DC image sequence constructed from the MPEG video sequence. Then we use the region segmentation-based projective histogram and its moments as database indices for video retrieval.