一种复杂背景下的手势分割新方法

A New Method of Hand Gesture Segmentation with Complex Backgrounds

从复杂的彩色图像中快速划分出手势,是手势识别的关键步骤,它处理的好坏直接影响到手势识别的效果。本文在分析现有方法的基础上,提出了一种基于复杂背景下的手势分割新方法。首先,在RGB模式空间里利用1.0<R/G<1.3建立第一个肤色模型,然后将HSV空间改进为更适合于肤色特征的模式空间HS謘V,在对H特征进行简化的基础上,建立第二个肤色模型,最后将这两种肤色模型相结合,得到复杂背景下的手势分割新方法。实验结果表明,这种方法的算法简单,易于实现,分割效果较好,并可用于实时图像处理。

To locate the image of the hand area currently is an important step in the hand gesture recognition based on the complex background. Its result has the great effect on the hand gesture recognition. A new method of hand segmentation with the complex background that realizes the hand segmentation on the base of the two skin color models, the one is RGB with the condition of 1.0〈R/G〈1.3 and the other is HS＇V, which is the improvement of the HSV is proposed, which has been proved that it is suitable for segmenting the hand even from the real-time color image, since its simple algorithm is easy to be realized.