基于高斯肤色模型的人脸区域及下巴检测  

The detection of face and chin based on Gaussian skin color model

作　　者：周利琴[1] 谷林 

机构地区：[1]西安工程大学计算机科学学院,陕西西安710048

出　　处：《西安工程大学学报》2015年第6期751-755,共5页Journal of Xi’an Polytechnic University

基　　金：西安丹比信息技术有限公司资助项目(2013kj-292)

摘　　要：通过对正面证件照进行色彩空间的转换,建立基于YCbCr色彩空间的高斯肤色模型,并利用基于肤色相似度和二值化处理的方法对肤色区域进行检测,从而区分开人体的肤色区域和非肤色区域;继而利用人脸特征区域尺寸的比例关系以及下巴阴影部分与下巴区域的亮度差别检测出下巴,区分开人脸区域和脖颈区域,最终准确标识出人脸区域.实验结果表明,经过对下巴的检测后,检测出的人脸区域更加准确.Through converting the positive identification photo into the color space of YCbCr, a distinction is made between the skin and non-skin regions by setting up a skin gaussian model based on the YCbCr color space, using the feature detection based on the skin color similarity and the binarization processing to detect the area of skin. Then the ratio of the face feature re- gion size,chin shaded part and the difference of the chin area brightness are used to distinguish face region and neck area. Eventually the face region is identified. The result proves that the face region is accurately identified by using the detection of chin.

关 键 词：高斯模型 色彩空间 人脸检测 下巴检测 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]