Tracking Camouflaged Objects with Weighted Region Consolidation

We propose an iterative method of Weighted Region Consolidation to track a camouflaged object within a background of very close colour in an image sequence, where the colours for both the object and the background are unsteady with large noises. More precisely, we will first detect the object motion based on both spatial and intensity densities by locating pixels with high motion probabilities from the difference data of successive frames. We then enhance the moving object iteratively, i.e. to consolidate the object region, by evaluating for each pixel its weighted overall neighbourhood intensity based on the pixel distances and intensity. The contour of the object’s moving area is then constructed, with an additional iterative process to refine the contour of the object in the current frame. We also propose to further refine the object contour, if needed, by the consideration of the neighbourhood intensities and shape similarity, the propagation of the initial contour, as well as the incorporation with other scattered and sketchy object boundaries obtained by another modified weight-based method applied on the edge map. Our proposed tracking scheme should be able to be eventually applied to monitoring the behavior of camouflaged wild life for their protection.