An Anti-Occlusion Correlation Filter Algorithm for Visual Tracking

Abstract

Visual tracking is one of the most challenging problems in computer vision, and occlusion is one of the difficulties in visual tracking. Since the appearance of the correlation filter tracker, it has attracted the attention of researchers with its superior speed and performance. But the correlation filter tracker has low robustness to occlusion problem. In order to improve the performance of the tracker, we design a dual-feature and dual-correlation filter based on Kernel Correlation Filter (KCF) by fusing spatio-temporal correlation filter, meanwhile, we propose a novel adaptive model updating method. Finally, the algorithm is tested on the OTB data benchmark. Experimental results show our result is better than many classical algorithms including KCF.