Video analysis for situation and threat recognition

Vision is one essential cue in present and future security systems since it provides comprehensive information of the surroundings. Especially in areas like surveillance and threat detection, vision provides indispensable information about the environment which allows detailed analysis. One important aspect in surveillance and threat detection is "person centric" analysis since people and their behavior are the focus of attention in many security systems. In this paper, we present a system for person centric video analysis. In particular, we introduce a system for person behavior recognition that is specifically suited for tasks like infrastructure protection and threat detection. For that, we combine state-of-the-art computer vision algorithms with high-level reasoning approaches.