手势识别中的肤色分割方法研究

Research on Skin Color Segmentation Method in Gesture Recognition

手势识别通常利用肤色特征进行手部定位,因此肤色分割直接影响到手势识别的准确率。为了提高手势识别的准确率,本文对常用于肤色分割的H、Cr、Cb、Cg颜色通道及其组合模型进行了肤色分割研究,并经过大量的仿真实验,发现基于HCrCg混合式颜色空间的肤色模型,不仅提高了对类肤色区域的抗干扰能力,还有效地减少光照条件的影响,可应用于手势识别中。

Gesture recognition usually uses skin color characteristics for hand positioning,so skin color segmentation directly affects the accuracy of gesture recognition.In this paper,the skin color segmentation of H,Cr,Cb,Cg color channels and their combination models commonly used in skin color segmentation is studied,and after a large number of simulation experiments,it is found that the skin color model based on HCrCg hybrid color space not only improves the anti-interference ability of skin color like areas but also effectively reduces the impact of lighting conditions,which can be applied to gesture recognition.