Dynamic video surveillance systems guided by domain ontologies

In this paper we describe how the knowledge related to a specific domain and the available visual analysis tools can be used to create dynamic visual analysis systems for video surveillance. Firstly, the knowledge is described in terms of application domain (types of objects, events... that can appear in such domain) and system capabilities (algorithms, detection procedures...) by using an existing ontology. Secondly, the ontology is integrated into a framework to create the visual analysis systems for each domain by inspecting the relations between the entities defined in the domain and system knowledge. Additionally, when necessary, analysis tools could be added or removed on-line. Experiments/Application of the framework show that the proposed approach for creating dynamic visual analysis systems is suitable for analyzing different video surveillance domains without decreasing the overall performance in terms of computational time or detection accuracy. (6 pages)