Novel Multi-window Tracking Based on Model Update

In clutter background, the performance of tracking target can be influenced by the factors such as illumination, camera angle. Meanwhile the target can be occluded by some obstacles in the background or be occluded by the target itself. To solve those problems, a multi-window tracking method is proposed, which represents the tracking target with several windows that each one corresponds with a tracker. There may exist overlaps among the windows or may contain the scene around the tracking target. And a model updated method is used to solve the mismatching between candidate model and reference model. Each tracker has an auxiliary model constructed on color statistical knowledge of the object. Experimental results clearly demonstrate the effectiveness of this method on self-occlusion problem or illumination influence in a complex background.