基于RGB-H-CbCr新颜色空间的肤色检测算法研究

The Study of Skin Detection Algorithms Based on the RGB-H-CbCr New Color Space

针对以往肤色检查算法很少考虑亮度,仅靠色度信息检测肤色,忽略了由于不同人种和亮度差异而导致的肤色像素在颜色空间上的分布不同,本文提出了一种基于RGB-H-CbCr新颜色空间的肤色检测算法。该算法综合了肤色像素在RGB、YCbCr和HSV三种颜色空间上的分布特征,构建出RGB-H-CbCr新颜色空间肤色模型,来辨别肤色和非肤色点。文中利用Matlab分别对不同人种的肤色图像进行实验仿真,实验结果表明该方法能有效检测人体肤色,能有效提高肤色检测的正检率。

The previous skin check algorithm seldom considered brightness.It only detected color of skin on color information,ignoring the differences of ethnic and racial differences and skin brightness which caused the differences of distribution of color pixel in color space.This paper puts forward a new color space skin detection algorithm based on the RGB-H-CbCr.This method combines the distribution characteristics of the skin color pixel in3 kinds of color space of RGB,YCbCr and HSV,and construct RGB-H-CbCr new color space skin model,to distinguish the color of skin and the non-color of skin point.This paper experimentally tests respectively different ethnic skin color image by using Matlab.The experimental results show that this method can effectively detects human body color of skin and can highly improve the effectiveness of detecting of the skin detection.