Abstract:

In this paper, we propose a fast object tracking algorithm based on template matching and region information fusion extraction. The method reduces the search candidates based on the prediction framework. Besides, the region information of the object is extracted to assist templatebased object tracking. The prediction model also reduces the computational complexity of object tracking tasks while the object region information improves the accuracy of object tracking tasks. The proposal combines the advantages of the object template and the object information extraction, thus improving the accuracy of object tracking. The tracking shift and object occlusion are handled  using the object motion information and object region information. Experimental results show that the proposed algorithm has robust performance when tracking drifting and blocked objects.

Key words: object tracking;information fusion;template matching;block analysis