A probabilistic framework for semantic indexing and retrieval in video

Abstract
This paper proposes a novel probabilistic framework forsemantic indexing and retrieval in digital video. The componentsof the framework are multijects and multinets. Multijectsare probabilistic multimedia objects [6] representingsemantic features or concepts. A multinet is a probabilisticnetwork of multijects which accounts for the interactionbetween concepts. The main contribution of this paper isa Bayesian multinet which enhances the detection probabilityof individual multijects,...

This publication has 12 references indexed in Scilit: