Abstract:

An automatic object tracking and video summarization method for multicamera systems with a large number of non-overlapping field-of-view cameras is explained. In this system, video sequences are stored for each object as opposed to storing a sequence for each camera. Object-based representation enables annotation of video segments, and extraction of content semantics for further analysis and summarization. Objects are tracked at each camera by background subtraction and mean-shift analysis. Then the correspondence of objects between different cameras is established by using a Bayesian Belief Network. This framework empowers the user to get a concise response to queries such as "which locations did an object visited on Monday and what did he do there?"