Camouflage modeling for moving object detection
Xiang Zhang, Ce Zhu · 2015
Discriminative feature based modeling (DFM) is widely used for moving object detection, which, however, may tend to fail when encountering camouflage problems. In this paper we propose a new strategy, camouflage modeling (CM), to detect camouflaged moving objects. In view that a camouflage area is highly content dependent of foreground and the nearby background information, we model both the background and camouflaged foreground respectively, and further identify the truely camouflaged areas. Finally, DFM and CM are fused to perform complete object detection. Experiments are conducted on testing sequences to demonstrate the effectiveness of the proposed method.