Surveillance Event Interpretation Using Generalized Stochastic Petri Nets

In this paper we present video event representation and recognition approaches that are based on Generalized Stochastic Petri Nets (GSPN). Along with the typical modeling capabilities of GSPN for video recognition, we propose to integrate the Petri net marking analysis for better scene understanding. This work focuses on behavior modeling and uses the results of an external module for object detection, tracking and classification. The proposed approach is evaluated using the developed surveillance system which can recognize events from videos and give a textual expression for the detected behavior. The experimental results illustrate the ability of the system to create complex spatiotemporal relations and to recognize the behavior of one or multiple objects in various video scenes.