Abstract:In this paper we present an effective and fast tracking algorithm, in which object tracking is achieved by solving L2-regularized least square (L2-RLS) problem-s within a Bayesian inference framework. Firstly, we model the appearance of the tracked target with P-CA basis vectors and square templates which make the tracker not only exploit the strength of sub space repre-senation but also explicitly take partial occlusion into consideration. Secondly, we adopt the L2-regularized least square method to solve the proposed representation model. Compared with the complicated LI-based algorithm, it provides a very fast performance without loss of accuracy in handling the tracking problem. In addition, a novel likelihood function and a refined update scheme are designed to further improve the robustness of our tracker. Both qualitative and quantitative evaluations on challenging videos demonstrate that the proposed method performs favorably against several state-of-the-art tracking algorithms.

Object tracking with L2-RLS