Abstract

In this paper, we propose a novel structural correlationfilter (SCF) model for robust visual tracking. The proposedSCF model takes part-based tracking strategies into ac-count in a correlation filter tracker, and exploits circularshifts of all parts for their motion modeling to preserve tar-get object structure. Compared with existing correlation fil-ter trackers, our proposed tracker has several advantages:(1) Due to the part strategy, the learned structural correla-tion filters are less sensitive to partial occlusion, and havecomputational efficiency and robustness. (2) The learnedfilters are able to not only distinguish the parts from thebackground as the traditional correlation filters, but also exploit the intrinsic relationship among local parts via spatial constraints to preserve object structure. (3) The learned correlation filters not only make most parts share similar motion, but also tolerate outlier parts that have different motion. Both qualitative and quantitative evaluations on challenging benchmark image sequences demonstrate thatthe proposed SCF tracking algorithm performs favorably a-gainst several state-of-the-art methods.