Real-time detection of removed objects based on multi-cue fusion
Weiping Sun, Shan Xu, Xiaoping Chen · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2009
In this paper, we propose a method to detect removed or stolen objects in complex environments robustly and efficiently based on the multi-cue adaptive fusion. In this method, the target observation is represented by multiple cues. When fusing each cue, a fusion scheme based on fuzzy logic has been developed. To solve the problem raised by part occlusions, target object is divided into multiple sub-parts and subsequently matching of at least one sub-part between current fields and initial model leads to a final decision that the target object is safe. Furthermore this method can distinguish target object into stolen or occluded objects by analyzing the matching degree of sub-parts with their initial models. Experimental results have demonstrated the effectiveness of the proposed method.