基于Kalman滤波器与肤色模型的手势跟踪方法

有效和鲁棒的手势跟踪是动态手势识别的前提，针对手势及其运动的特点，提出了结合Kalman滤波器和肤色模型的手势运动目标跟踪方法.首先通过背景差法和YCb’Cr’空间上的椭圆肤色模型检测出手部运动目标，通过目标区域的空间结构参数来设置Kalman滤波器的各项运动参数，然后计算空间结构特征的跟踪匹配函数对目标预测位置进行修正，获得运动手势目标区域及其运动轨迹.实验结果表明，所提方法能有效地跟踪手势，并能较好地适应手势在运动过程中的手形变化、轨迹转弯等情况，检测准确，鲁棒性高.

Hand Gesture Tracking Method Based on Kalman Filter and Skin Color Feature

The effective and robust hand gesture tracking is the premise of the dynamic hand gesture recognition. According to the characteristics of hand gesture movement, a hand gesture tracing method was presented on the basis of the Kalman filter and the skin color model. Firstly, the hand movement target was recognized with the background difference and the skin color model. At the same time, the movement parameters of Kalman filter were set by the parameters of spatial structure. Then, the tracking matching function based on the feature of spatial structure was calculated to correct the target predicted position, and the target region of dynamic hand gesture and its motion track were also got. The results showed that the proposed method was effective, accurate and robust to track hand gesture and was adapted to hand change, turning path and other situations.