在线矩形特征选择的压缩跟踪算法

针对压缩跟踪算法无法选择合适的矩形特征,易出现目标漂移、丢失现象,提出了一种基于在线矩形特征选择的压缩跟踪算法。首先,在初始化阶段生成投影矩阵,利用该投影矩阵提取特征来构造候选特征池,在特征池中使用矩形特征来表示目标特性,并去除与目标差异较大的矩形特征,最后计算分类分数最大的窗口,并将其作为目标窗口,从而实现跟踪。实验结果表明,该算法特征总数量比压缩跟踪算法特征总数量减少了13%,且跟踪精度和鲁棒性方面得到了改善,对于320pixel×240pixel大小的视频平均处理帧速为20frame/s,满足实时性要求。

Compression Tracking Algorithm for Online Rectangle Feature Selection

Compressive tracking algorithm can not select appropriate object futures which will result in drifting or make tracking not accurate when the object is occluded or its appearance changes.To address this problem,this paper proposed a real-time compressive tracking algorithm based on rectangle feature selection.Firstly,generate projection matrixes are generated in an initial phase.And the projection matrixes are used to extract the feature to construct a feature pool.The rectangle feature is used to represent the characteristics of target in the feature pool, and the rectangular features with greater difference from the target characteristics are removed.Finally,the classifier is taken to process candidate samples by Bayes classification and response results to the classifier are taken as tracking results.The experimental results show that the proposed algorithm is about 13% lower than that of compressive tracking.It improves the trac-king accuracy and robustness,and the processing frame rate is 20 frame/s on a 320pixel×240pixel video sequence,which meets the requirements of real-time tracking.