Blurred target tracking based on sparse representation of online updated templates
Xiaofen Xing, Nanhai Zhang, Kailing Guo, Chunmei Qing, Jiali Deng, Huiping Qin · 2016
Motion blur is pervasive in object tracking due to the camera and target movement. Most approaches are prone to drift when the target is blurred. Sparse representation of both normal and blur templates are able to improve the robustness of the appearance model, but how to update the template set remains a difficult problem. To solve this problem, we propose a two-step observation correction strategy for template updating by: 1) treating motion blur as trivial information with Laplace distribution and using different ways to correct the normal target and the blur target; 2) utilizing incremental principle component analysis (PCA) to update the normal template set. Experiments on challenging videos show the proposed algorithm outperforms several state-of-the-art methods.