Adversarial behavior recognition from layered and persistent sensing systems
Georgiy Levchuk, Olga L. Mendoza-Schrock, Wayne L. Shebilske · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2009
Advances in multi-sensor automated target recognition, tracking, and Wide Area Persistent Surveillance promise to enable a broad spectrum of intent and behavior recognition models. However, a significant gap remains between coordinated behavior analyses tools working at high information abstraction and target identification and tracking systems working with direct inputs from motion imagery. In this paper, we describe the problem of modeling adversarial behavior signatures faced by Air Force researchers, present a range of solutions to automate the discovery of behavior patterns, and outline the gap in the research space to enable efficient integration of multi-level information exploitation, analyses and sensor management tools.