SEGMENTING MOVING OBJECTS: THE Modest VIDEO OBJECT KERNEL

A system separating objects moving within a slow changing background is presented. The originality of the approach resides in two related components. First, the change detection robust to camera noise which does not require any sophisticated parametric tuning as it is based on a probabilistic method. Second, the change is detected between a video frame representing a scene at a given time, and reference that is updated continuously to take into account slow variation in the background. The system is particularly suitable for indoor and outdoor surveillance. Simulation results show that the proposed scheme performs rather well in extracting video objects, with stability and good accuracy, while being of a relatively reduced complexity.