A Query Answering Surveillance System for Detecting and Tracking Moving Objects using Bounding Rectangles

In many computer vision applications such as surveillance systems, detecting moving objects is the first step of information extraction. This paper introduces a query answering system that aims at responding to user queries without the need to scan the whole video tape. The proposed system is based on finding the minimum boundary rectangles (MBRs) of moving objects which appear in the video. It detects moving objects in the video frame set, keeps track of those objects, and saves information to answer future