YCgCr颜色空间的肤色聚类人脸检测法

Face detecting based on skin color clustering in a color space YCgCr

由于YCgCr颜色空间好的聚类性。提出一种在YCgCr颜色空间的人脸检测,首先肤色在CgCr颜色空间进行投影,对像素点区域进行拟合,初步去除图像中的非肤色,得到肤色第一次分割的图像,实验证明这种拟合可以起到好的去除非肤色像素的作用。然后把YCgCr颜色空间作为肤色模型建立彩色空间,用高斯模型在已提取的图像上进行人脸检测。实验证明在对CgCr空间拟合的基础上,再运用高斯模型进行人脸检测,该算法提高了人脸检测的速度,并且有较高的检测率和较低的漏检率。

Due to the clustering is excellent in a color space YCgCr.This paper provides a way of face detection in a color space YCgCr.Firstly,skin colors are projected in a color space CgCr and the fitting of distribution is carried through in order to wipe off a part of non skin color and gain the intersected image for the first time.Experimental results indicate that this fitting of distribution can reduce the great mass of processing pixels of non-skin color.Then this paper puts Color Space YCgCr as the color model to build a color space and use Gauss model to detect human face on a base of the intersected image.Experimental results indicate that Gauss model to detect human face on a base of fitting of distribution in CgCr color space can improve the speed of face detection and have upper detecting ratio.This way also has upper detection ratio and lower ratio of missing detection and low ratio of false detection.