计算机应用 ›› 2012, Vol. 32 ›› Issue (05): 1404-1406.

• 信息安全 • 上一篇    下一篇

基于对称核主成分分析的人脸识别

刘嵩1,2,罗敏1,张国平2,3   

  1. 1. 湖北民族学院 信息工程学院,湖北 恩施 445000
    2. 华中师范大学 物理科学与技术学院,武汉 430079
    3. 华中师范大学物理学院
  • 收稿日期:2011-10-31 修回日期:2011-12-18 发布日期:2012-05-01 出版日期:2012-05-01
  • 通讯作者: 刘嵩
  • 作者简介:刘嵩(1979-),男,湖北孝感人,讲师,博士研究生,CCF会员,主要研究方向:模式识别、嵌入式系统设计;罗敏(1978-),女,湖北随州人,讲师,主要研究方向:光电子技术;张国平(1969-),男,湖北汉川人,教授,博士生导师,主要研究方向:模式识别、光电子技术。
  • 基金资助:

    湖北省自然科学基金资助项目(2009CDB069)

Face recognition based on symmetrical kernel principal component analysis

LIU Song1,2,LUO Min1,ZHANG Guo-ping2,3   

  1. 1. College of Information Engineering, Hubei Institute for Nationalities, Enshi Hubei 445000, China
    2. College of Physical Science and Technology, Huazhong Normal University, Wuhan Hubei 430079, China
    3.
  • Received:2011-10-31 Revised:2011-12-18 Online:2012-05-01 Published:2012-05-01
  • Contact: LIU Song

摘要: 为了提高人脸识别技术的实用性,结合人脸镜像对称性和核主成分分析提出了一种新的人脸识别方法。首先利用小波变换压缩人脸图像数据,获取小波分解的低频分量,再通过镜像变换得到镜像偶对称图像和镜像奇对称图像,然后分别对奇偶对称图像进行核主成分分析提取奇偶特征,并且通过加权因子对奇偶特征进行融合,最后采用最近邻分类器分类。基于ORL人脸数据库的实验结果表明:该算法增大了样本容量,在一定程度上克服了光照、姿态的不利因素,提高了人脸识别率。

关键词: 人脸识别, 镜像对称, 特征提取, 核主成分分析, 最近邻分类器

Abstract: In order to improve the practicability of face recognition technology, a new face recognition method was proposed by adopting the facial mirror symmetry and Kernel Principle Component Analysis (KPCA). Firstly, the original images were decomposed by wavelet transform, and the low frequency components could be obtained. Then, the odd symmetry samples and the even symmetry samples were obtained by mirror transforming. Odd/even eigen vector were separately extracted through KPCA and fused to composite features by an odd-even weighted factor. A nearest neighbor classifier was used to classify the images. The proposed method was tested on the ORL face image database. The experimental results show the method can increase the sample capacity, overcome the effect of illumination and posture, and raise the recognition rate. Besides, in the comprehensive performance, it is better than contrast method.

Key words: face recognition, mirror symmetry, feature extraction, Kernel Principle Component Analysis (KPCA), nearest neighbor classifier

中图分类号: