Particle Filter Optimized by CamShift for Scale Adaptive Face Tracking

摘要：

In this context, a novel algorithm is proposed for scale adaptive face tracking. Mean-Shift process is applied to Particle Filter (PF) framework, where PF is optimized by MeanShift procedure due to its property of accurate tracking and fast convergence. And the accuracy position determination makes window size computation possible in the proposed method. On the other hand, the algorithm maintains multihypothesis tracking, so it is still robust. Additionally, an effect mechanism is proposed for checking tracking results, which provides useful information for next frame, in this paper, we have described and evaluated the performance of the proposed algorithm. The experimental results verify its efficiency and robustness for face tracking.

关键词： face tracking particle filter Mean-Shift histogram scale adaptive

作者: Xinmei Li Yi Cao Jun Kong Jin Zhang Danni Yang

作者单位: Computer School, Northeast Normal University Changchun, China Computer School, Northeast Normal University Key Laboratory for Applied Statistics of MOE Changchun,

会议类型: 国际会议

会议名称: The 10th International Conference on Intelligent Technologies(第十届智慧科技国际会议 InTech09)

会议地点: 桂林

会议语种:英文

页码: 578-583

在线出版日期: 2009-12-12（万方平台首次上网日期，不代表论文的发表时间）