Mosaic-based clustering of scene locations in videos

We present an approach for compact video summaries that allows fast and direct access to video data. The video is segmented into shots and scenes using a previously proposed method, then motion analysis is used to select representative shots for each scene. In contrast to approaches to video indexing which are based on key-frames, we use mosaics constructed from the representative shots for an efficient mosaic-based scene representation. We use a novel method for mosaic comparison to spatially cluster scenes and create a highly compact non-temporal representation of video. Our scene-based representation allows accurate comparison of scenes across different video data, and serves as a basis for indexing of whole video sequences