ONTOLOGY AND BAYESIAN NETWORK BASED AUTOMATIC VIDEO ANNOTATION
Abstract
In this paper 0e present an automatic video annotation s2stem, based on a domain ontolog2, 0hich contains the le1icon used for the annotation process. A Ba2esian Net0ork is also used to recogni3e events and scenarios in the video fragments. An e1perimental testing 0as conducted using the CAVIAR test video dataset