基于最大池图匹配的形变目标跟踪方法

该文提出了一种基于最大池图匹配的形变目标跟踪算法,适用于跟踪目标产生较大形变或者严重遮挡等场合.此方法首先将目标搜索区域过分割为候选目标部件并建立动态图表示,即目标部件的表象特征和它们之间的几何位置关系.然后采用最大池图匹配算法,得到目标图和候选图中部件的匹配关系,从而确定出目标位置的置信图.联合考虑目标整体和目标部件对目标位置的支持,投票决定出精确的目标位置.在各种形变目标的跟踪序列测试下,该算法与其他跟踪器的对比验证了其有效性和鲁棒性.

Deformable Object Tracking Based on Max-pooling Graph Matching

This paper develops a novel deformable object tracking algorithm based on max-pooling graph matching,which can be applied in the scenes with large deformations and severe occlusions.The dynamic graph is built based on candidate parts extracted by over-segmentation method from searching area,namely feature representation of candidate parts and geometric structure between them.Based on max-pooling graph matching method,the matching relations between target parts and candidate parts are found to calculate the confidence map of target location.Considering both the support of holistic target and local parts,the optimal target location can be determined.Compared to state-of-the-art methods,experimental results on several deformable sequences demonstrate the effectiveness and robustness of the proposed method.