基于高斯肤色模型与Blob算法的人脸检测  

Face Detection Based on Gaussian Color Model and Blob Algorithm

作　　者：李岚 张云 LI Lan;ZHANG Yun(School of Digital Medial,Lanzhou University of Arts and Science,Lanzhou 730000,Gansu,China)

机构地区：[1]兰州文理学院数字媒体学院,甘肃兰州730000

出　　处：《咸阳师范学院学报》2018年第4期60-63,共4页Journal of Xianyang Normal University

基　　金：甘肃省高校科研项目(2015B-133;2016A-105);兰州文理学院科研项目(2013YBTS02)

摘　　要：针对彩色图像中背景复杂的人脸检测率低的问题,提出了一种基于肤色模型的人脸检测方法。在新颜色空间YCgCr上对图像进行肤色检测并应用预处理技术缩小人脸检测的搜索区域;在候选区域的基础上,应用Blob分析方法确定任意旋转角度的人脸,去除非人脸区城,降低误检率。实验结果表明:文中方法对不同尺寸、任意旋转的人脸有较好的检测效果。Aiming at the low face detection rate in color image of complex background, a face detection method was proposed based on color model. Firstly, the color was detected on the new color space YCgCr, and the search area was reduced by applying the processing technique. Then, on the basis of the candidate region, the Blob analysis method was used to determine the face of any rotation angle with non-human face area removed, reducing the false detection rate. The experimental results show that this method has better effect on different sizes and arbitrary rotation faces.

关 键 词：人脸检测 人脸姿态 YCGCR颜色空间 BLOB分析 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]