基于子空间联合模型的视觉跟踪

VISUAL TRACKING BASED ON SUBSPACE COLLABORATIVE MODEL

目标跟踪是计算机视觉的重要组成部分,其鲁棒性一直受到目标遮挡,光照变化,目标姿态变化等因素的制约。针对这个问题,提出了基于子空间联合模型的视觉跟踪算法。算法为了克服遮挡对目标跟踪的影响,采用局部动态稀疏表示进行遮挡检测,根据遮挡检测结果来修正增量子空间误差。此外,在稀疏子空间基础上计算目标模板和候选模板的相似性。在粒子滤波框架下,联合候选目标增量误差和相似性实现目标跟踪。通过在多个具有挑战性的视频序列上进行实验,表明该算法具有较好的鲁棒性。

Target tracking is an important part of computer vision, and its robustness is always restricted to target occlusion, illumination variation and target pose change and so on. To this end, this paper proposes a visual trackingalgorithm based on subspace collaborative model. In order to overcome the influence of occlusion on target tracking, this algorithm rectifies incremental subspace error by result of occlusion detection using local dynamic sparse representation.Besides, the similarity between target template and candidate template is computed based on local dynamic sparse representation. In the framework of particle filter, this algorithm is achieved based on combining incremental error withsimilarity. The experimental results on several sequences show that this algorithm has better performance of tracking.