Object tracking algorithm using objectness detection

To solve the tracking drift problem caused by the low discrimination of object appearance information in complex environment, the paper proposes an object tracking algorithm using objectness detection. First, the algorithm obtains the preliminary object prediction state with kernelized correlation filters. Then, according to the objectness detection principle of the proposal bounding box, it generates the original proposal bounding box set with the consideration of the scale and aspect ratio, and further gets optimized sets with the refined selection criterion. By introducing motion continuity, the prediction location and scale based on the proposal bounding box are calculated, and then the final optimum object state estimation is acquired comprehensively. Finally, taking into account the occlusion influence judge of target appearance at the current frame, the corresponding template updating scheme is given. Experimental results demonstrate that the novel algorithm achieves robust tracking performance in various typical testing scenarios.