Abstract
The transition from analog to digital video enables the consumer access to an enormous amount of high quality video material. In this paper, we describe architecture of a video content filtering system through video analysis and abstraction. We explain the algorithms for video characterization, their performance and practical considerations for including them in the architecture for video filtering. The algorithms and the architecture for abstracted video representation can be used in a variety of consumer domain applications, such as a personalized news retrieval system, home video library or commercial skipping.