Transparent Access to Video Over the Web: A Review of Current Approaches

Abstract

This chapter is a technical review of computer tools which aim to provide manageable computer-based representations of video sequences. The paper is part of a series of publications (see Macer and Thomas 1996a,b; Macer et a1. 1996; Macer and Thomas 1999a–e) which describe work in the development of solutions to the problems of accessing video sequences on computer, particularly over the Web. Several approaches are reviewed here, which range in approach from those which use a highly abstract representation, dissimilar in nature from the medium of video and which rely on conventional, rigid information retrieval techniques such as database querying, to those in which the representation is similar in nature to the video which it represents — for example images taken from the video sequence itself and which use less traditional, more ad hoc retrieval techniques. The chapter is intended to provide a comprehensive, coherent, critical review of those approaches which are based on abstract representations using conventional retrieval techniques, and those which employ more literal representations using newer retrieval techniques. The systems discussed are introduced in an approximate order of decreasing abstraction and increasingly ad hoc retrieval techniques.