| 基于肤色特征和动态聚类的彩色人脸检测 |
| |
| 引用本文： | 何光宏,潘英俊,吴芳.基于肤色特征和动态聚类的彩色人脸检测[J].光电工程,2004,31(11):47-50. |
| |
| 作者姓名： | 何光宏 潘英俊 吴芳 |
| |
| 作者单位： | 1. 重庆大学数理学院,重庆,400044
2. 重庆大学光电工程学院,重庆,400044 |
| |
| 基金项目： | 重庆大学基础与应用基础研究基金(717411062) |
| |
| 摘 要： | 在人类视觉机制和肤色聚类特性的基础上,提出了一种复杂背景下人脸检测方法。该方法采用K-均值动态聚类分析算法,利用人类肤色特征在输入图像中检测包含人脸的似人脸区作为候选人脸,再用同样的方法对候选人脸区域进行扫描,得到真正的人脸。实验结果表明,该方法的正确检出率达到84%,受背景、光照、角度、姿态的影响很小,具有较好的鲁棒性。
|
| 关 键 词： | 图像检测 人脸检测 肤色特征 聚类特性 |
| 文章编号： | 1003-501X(2004)11-0047-04 |
| 收稿时间： | 2004/1/5 |
| 修稿时间： | 2004年1月5日 |
Colorful human face detection based on skin color features and dynamic clustering |
| |
| HE Guang-hong,PAN Ying-jun,WU Fang.Colorful human face detection based on skin color features and dynamic clustering[J].Opto-Electronic Engineering,2004,31(11):47-50. |
| |
| Authors: | HE Guang-hong PAN Ying-jun WU Fang |
| |
| Affiliation: | HE Guang-hong1,PAN Ying-jun2,WU Fang2 |
| |
| Abstract: | A method for human face detection under complex background is proposed on the basis of human vision mechanism and skin color clustering features. The method extracts the candidate human face area which contains human face and semblance human face from the input images by means of K-means dynamic clustering analysis algorithm and human skin color features. The candidate human face area will be scanned by the same method and the real human face can be obtained. The experimental results show that the correct detection ratio of the method can be up to 84%. The influence of background, light illumination, angle and altitude on detection is very small and the method has good robustness. |
| |
| Keywords: | Image detection Human face detection Skin color features Clustering features |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
