中文摘要：

稀疏表示在目标跟踪中已取得了良好的跟踪效果,但过完备词典模式单一、数据庞大,稀疏系数需用复杂的优化算法求解,会限制此类算法的跟踪性能提高。因此,本文在粒子滤波框架下,提出了一种基于多模态词典的目标跟踪算法。首先,创建长、短周期的正样本模板,结合负样本模板共同构成多模态词典,用以表征采样目标当前状态;其次,根据样本与词典之间的多模态相关系数,对目标进行粗跟踪,得到候选跟踪结果;最后,利用Local Maximal Occurrence（LOMO）特征构建候选跟踪目标与多模态词典的观测似然函数,取具有最大似然度的候选跟踪目标作为最终的跟踪结果。实验结果表明,本文算法在遮挡、光照变化和背景干扰情况下均具有较强的跟踪鲁棒性。

英文摘要：

Sparse representation-based methods have been successfully applied to visual tracking.However,the over complete dictionary mode is single and data is large,and the sparse coefficients need to be solved by the complex optimization algorithm,which will limit their tracking performances.In this paper,within the tracking framework of particle filter,we propose a tracking method based on the multi-modality dictionary learning.Firstly,a long and short period of object templates are created,combined with background templates to form a multi-modality dictionary to characterize the current state of sampled object.Secondly,according to the multi-modal coefficients between the sampled objects and the dictionary,the target is tracked roughly with the candidate tracking results obtained.Finally,the observation likelihood functions of the candidate tracking results and the multi-modality dictionary are constructed by using LOMO features,and the candiadate tracking result with the maximum likelihood is taken as the final tracking result.Experimental results demonstrate that the proposed method has strong tracking robustness in the case of occlusion,illumination change and background interference.