Experiential Sampling for video surveillance

Due to the decreasing costs and increasing miniaturization of video cameras, the use of digital video based surveillance as a tool for real-time monitoring is rapidly increasing. In this paper, we present a new methodology for real-time video surveillance based on Experiential Sampling. We use this framework to dynamically model the evolving attention in order to perform efficient monitoring. We exploit the context and past experience information in order to detect and track moving objects in surveillance videos. Moreover, we take the situation of multiple surveillance cameras into account and utilize the experiential sampling technique to decide which surveillance video stream to be displayed on the main monitor. This can tremendously help in reducing the manual operator fatigue for multiple monitor situation. We have implemented the developed algorithms and experimental results have been presented to illustrate the utility of the proposed technique.