Video indexing by spatial representation

Abstract

Recent approaches to video indexing and retrieval are either pixel oriented or object oriented. While the former focus on motion and changes therein, the latter focus on spatial relations among objects in the scene.In this paper, a spatial knowledge representation combining both approaches is proposed. This representation supplements spatial knowledge of visual objects with the information of their pixel positions in the video frame. It will provide a practical way t o construct video indices, enabling searching for and retrieval of video sequences that contain motion as well as sparsely disjointed objects.