Online object tracking by sparse and structural model

The objects tracking often meets the phenomenon, such as part or heavy occlusion, illumination variation, clutter background and scale variation and so on, during the tracking process, and which are the main challenge for object tracking. In this work, an online object tracking algorithm using local structural model with overlapped patches is proposed, the target is represented local structural feature with sparsity, furthermore, the sparse coefficients are pooled by the L2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$L_{2}$$\end{document}-pooling method, and the target can be located more accurately and the occlusion problem can be better handled with this strategy, In addition, the paper develops an adaptive updating strategy based on increment subspace learning and sparse representation, this not only helps weaken the influence of illumination but also reduces the possibility of drifting. Numerous experiments demonstrate that the proposed algorithm performs more robustly and effectively against several state-of-the-art algorithms.