一种基于Contourlet变换的人脸识别方法  

Face Recognition based on Contourlet Transform

作　　者：夏思宇[1] 李久贤[1] 袁晓辉[1] 夏良正[1] 

机构地区：[1]东南大学自动化学院,南京210096

出　　处：《信号处理》2008年第4期631-634,共4页Journal of Signal Processing

基　　金：浙江省长三角联合攻关项目(2005E60007)

摘　　要：Contourlet变换是一种新的多尺度几何分析方法,它不仅具有小波变换的多分辨率特性和时频局域特性,还具有很强的方向性和各向异性。该文分析了Contourlet变换的基本原理与变换的特点,提出了基于Contourlet变换的人脸识别方法,分别提取Contourlet变换的低频系数特征与高频方向子带统计特征进行实验。实验结果证明,Contourlet变换的低频系数特征具有优异的识别特性,而其高频方向子带统计特征则刻画了人脸轮廓与局部器官形状等信息,也具有一定的识别性能。Contourlet is a new image representation scheme which not only posses the main features of wavelets （ namely, multiscale and time-frequency localization） , but also have directionality and anisotropy. In this paper, the characteristic of contourlet transform is analyzed and a method based on contourlet transform for face recognition is proposed. In this method, the features of the low frequency and directional subbands in contourlet transform are extracted respectively. The experimental results show that the coefficients of the low frequency of contourlet transform achieve higher recognition rates and the statistical feature of directional subbands is also helpful for face recognition.

关 键 词：人脸识别 小波变换 CONTOURLET变换 

分 类 号：TN911[电子电信—通信与信息系统] TP391.41[电子电信—信息与通信工程]