计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (18): 1-3.DOI: 10.3778/j.issn.1002-8331.2010.18.001

• 博士论坛 • 上一篇    下一篇

肤色相似度和动态阈值相结合的肤色分割技术

郭 耸1,顾国昌1,蔡则苏2,刘海波1,沈 晶1   

  1. 1.哈尔滨工程大学 计算机科学与技术学院,哈尔滨 150001
    2.哈尔滨工业大学 计算机科学与技术学院,哈尔滨 150001
  • 收稿日期:2010-03-25 修回日期:2010-05-04 出版日期:2010-06-21 发布日期:2010-06-21
  • 通讯作者: 郭 耸

Skin segmentation using similarity of skin color and dynamic threshold

GUO Song1,GU Guo-chang1,CAI Ze-su2,LIU Hai-bo1,SHEN Jing1   

  1. 1.School of Computer Science and Technology,Harbin Engineering University,Harbin 150001,China
    2.School of Computer Science and Technology,Harbin Institute of Technology,Harbin 150001,China
  • Received:2010-03-25 Revised:2010-05-04 Online:2010-06-21 Published:2010-06-21
  • Contact: GUO Song

摘要:

为了提高人脸检测的速度与性能,提出一种基于肤色相似度与动态阈值相结合的肤色分割方法。首先在YCgCr颜色空间计算肤色相似度,然后给出一种基于类间方差和类内离散度相结合的动态阈值确定方法,根据求得的动态阈值进行肤色分割,并对肤色分割后的二值图像进行滤除噪声处理。实验结果表明该方法改善了肤色分割性能,能够在复杂背景下实现肤色区域的精确分割,从而提高了人脸检测的速度和性能。

关键词: 人脸检测, 肤色分割, 肤色相似度, 阈值分割, 动态阈值确定

Abstract: A method for skin segmentation based on similarity of skin color and dynamic threshold is proposed to improve the speed and performance of face detection.Firstly,the similarity of skin color is calculated in YCgCr space.Then,the method based on between-cluster variance and within-class scatter for selecting the threshold dynamically is proposed.Skin color segmentation is implemented according to the dynamic threshold.The binary image obtained by skin segmentation is processed in order to eliminate noise.The experimental results show that the proposed algorithm improves the performance of skin segmentation,and can segment skin regions exactly in the complex background so as to improve the speed and performance of face detection.

Key words: face detection, skin segmentation, similarity of skin color, threshold segmentation, dynamic threshold selecting

中图分类号: