基于mean-shift算法的人脸实时跟踪方法*

复杂环境下实时人脸跟踪方法在视觉监控系统中具有很重要的意义，但目前的跟踪算法普遍存在目标遮挡、尺寸变化等过于敏感的不足，限制了其应用范围。提出了一种人脸检测、mean-shift算法与卡尔曼滤波器相结合的实时全自动人脸跟踪算法。实验结果表明该算法实时性很强，可以实现对运动人脸的快速跟踪，同时对目标遮挡也有很好的鲁棒性。

Real-time algorithm for face tracking based on mean-shift

A real-time face tracking under complex environment has significance in visual surveillance system, but one common problem to face tracking algorithms at present is their sensitivity to temporal occlusion and scale variations, which restrict them wide application. This paper proposed a real-time automatic face tracking algorithm which combined face detection and mean-shift algorithm with Kalman filter. In the experiments, the results show the proposed algorithm can track fast moving face successfully and have better robustness of occlusion.