Realtime object-of-interest tracking by learning Composite Patch-based Templates

In this paper, we propose a patch-based object tracking algorithm which provides both good enough robustness and computational efficiency. Our algorithm learns and maintains Composite Patch-based Templates (CPT) of the tracking target. Each composite template employs HOG, CS-LBP, and color histogram to represent the local statistics of edges, texture and flatness. The CPT model is initially established by maximizing the discriminability of the composite templates given the first frame, and automatically updated on-line by adding new effective composite patches and deleting old invalid ones. The inference of the target location is achieved by matching each composite template across frames. By this means the proposed algorithm can effectively track targets with partial occlusions or significant appearance variations. Experimental results demonstrate that the proposed algorithm outperforms both MIL and Ensemble Tracking algorithms.