基于YCbCr空间肤色自适应分割方法  

作　　者：谢杰 张键 付天豪 

机构地区：[1]中国电子科技集团公司第五十八研究所,江苏无锡214072 [2]中科芯集成电路有限公司,江苏无锡214072

出　　处：《科技创新与应用》2021年第26期132-136,共5页Technology Innovation and Application

基　　金：江苏省科技成果转化高性能实时微处理器系列产品研发及产业化(编号:BA2019012)。

摘　　要：针对现有人体肤色区域分割方法在变化的光照、复杂的背景环境下鲁棒性差、精确度不高的问题,提出一种肤色自适应分割方法。首先将图片转化到改进的YCbCr颜色空间获取待分割肤色图,在变化的光照下获取肤色像素信息。然后在YCbCr颜色空间的(Cb-Cr)空间建立椭圆边界肤色模型,并建立(Cr-Cb)肤色灰度直方图,在二维灰度直方图上确定肤色大致位置,过滤掉大部分背景干扰。最后在灰度投影直方图上用高斯函数拟合得到精确的动态肤色范围,去除类肤色干扰。实验结果表明,该算法能够在变化的光照下对肤色进行自适应分割,在复杂背景下也有较好的分割效果。In order to solve the problems of poor robustness and low accuracy of existing human skin color region segmentation methods in changing illumination and complex background environment,an adaptive skin color segmentation method is proposed.Firstly,the image is converted to the improved YCbCr color space to obtain the skin color image to be segmented,and the skin color pixel information is obtained under the changing illumination.Then the elliptical boundary skin color model is established in the(Cb-Cr)space of the YCbCr color space,and the(Cr-Cb)skin gray histogram is established.The approximate position of the skin color is determined on the two-dimensional gray histogram,and most of the background interference is filtered.Finally,the accurate dynamic skin color range is obtained by Gaussian function fitting on the gray projection histogram,and the skin-like interference is removed.The experimental results show that the algorithm can segment skin adaptively under changing illumination and has a good segmentation effect under complex background.

关 键 词：肤色分割 椭圆边界肤色模型 YCBCR颜色空间 图像检测 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]