Robust Object Tracking using Probabilistic Hypergraph Ranking and Superpixels

Abstract

Keywords: Visual tracking, transductive learning, hypergraph ranking, superpixel. Abstract. Online object tracking is a challenging issue because the appearance of an object tends to change due to intrinsic or extrinsic factors. In this study, we propose a robust tracking algorithm based on probabilistic hypergraph ranking and superpixels. The probabilistic hypergraph is constructed by mid-level visual cues and their spatial relationships. Then, the confidence map at mid-level cues is obtained by hypergraph ranking analysis, which takes the high order intrinsic relationships of superpixels into account. Third, Object tracking is formulated as a transductive learning issue, and the optimal target location is determined by maximum a posterior estimation on the ranking scores. Finally, a dynamic updating scheme is proposed to address appearance variations and alleviate tracking drift. A series of experiments and evaluations on various challenging sequences are performed, and the results show that the proposed algorithm performs favorably against other existing state-of-the-art methods.