A Scenario-Based Video Surveillance Data Modeling and Querying System

Automated video surveillance has emerged as a trendy application domain in recent years, and accessing the semantic content of surveillance video has become a challenging research area. The results of a considerable amount of research dealing with automated access to video surveillance have appeared in the literature. However, event models and content-based access to surveillance video have significant semantic gaps remaining unfilled. In this paper, we propose a scenario-based querying and retrieval model for video surveillance archives. In our model, a scenario is specified as a sequence of event predicates, that can be enriched with object-based low-level features. Our model also provides support for inverse querying, which can be considered as a tool for after-the-fact type of activity analysis. We have devised a visual query specification interface to ease the scenario-based query specification process. It is shown through performance experiments that our system has an effective expressive power and satisfactory retrieval accuracy in video surveillance. This work is supported in part by European Commission 6 Framework Program MUSCLE Network of Excellence Project with grant number FP6-507752, and Scientific and Technical Research Council of Turkey (TUBITAK) with grant number EEEAG-105E065. November 14, 2007 DRAFT