Locally discriminative stable model for visual tracking with clustering and principle component analysis

• Published in: IET computer vision Vol. 7; no. 3; pp. 151 - 162
• Main Authors: Zhang, Canlong, Jing, Zhongliang, Tang, Yanping, Jin, Bo, Xiao, Gang
• Format: Journal Article
• Language: English
• Published: Stevenage The Institution of Engineering and Technology 01.06.2013; Institution of Engineering and Technology; John Wiley & Sons, Inc; Wiley
• Subjects: Applied sciences; Clustering; Computer vision; discriminative stable regions; DSR; Entropy; Exact sciences and technology; hierarchical clustering; image motion analysis; Information, signal and communications theory; intrinsic appearance variations; locally discriminative stable model; maximal local entropy criterion; motion noise; object tracking; occlusions; pattern clustering; principal component analysis; principle component analysis; Signal and communications theory; Signal representation. Spectral analysis; Signal, noise; spatial discrimination; Target tracking; Tasks; Telecommunications and information theory; Tracking; Visual; visual tracking approach; discriminative stable regions; motion noise; DSR; visual tracking approach; spatial discrimination; maximal local entropy criterion; image motion analysis; locally discriminative stable model; pattern clustering; principle component analysis; target tracking; occlusions; object tracking; hierarchical clustering; principal component analysis; intrinsic appearance variations; Performance evaluation; Statistical method; Discriminant analysis; Automatic classification; Target tracking; Hierarchical classification; Tracking task; Entropy; Signal classification; Discrimination task; Principal component analysis
• ISSN: 1751-9632; 1751-9640; 1751-9640
• DOI: 10.1049/iet-cvi.2012.0180

The challenge of visual tracking mainly comes from intrinsic appearance variations of the target and extrinsic environment changes around the target in a long duration, so the tracker that can simultaneously tolerate these variabilities is largely expected. In this study, the authors propose a new tracking approach based on discriminative stable regions (DSRs). The DSRs are obtained based on the criterion of maximal local entropy and spatial discrimination, which enables the tracker to handle well distractors and appearance variations. The collaborative tracking incorporated hierarchical clustering can tolerate motion noise and occlusions. In addition, as an efficient tool, the principle component analysis is used to discover the potential affine relation between DSR and the target, which timely adapts to the shape deformation of the target. Extensive experiments show that the proposed method achieves superior performance in many challenging target tracking tasks.