Modeling Video Data for Content Based Queries: Extending the DISIMA Image Data Model