最大散度差鉴别分析及人脸识别  

Maximum Scatter Difference Discriminant Analysis and Face Recognition

作　　者：刘永俊[1] 陈才扣[1] 

机构地区：[1]扬州大学计算机科学与工程系,江苏扬州225009

出　　处：《计算机工程与应用》2006年第34期208-210,227,共4页Computer Engineering and Applications

基　　金：江苏省高校自然科学基金资助项目(05KJB520152);江苏省博士后科研计划资助项目。

摘　　要：传统的Fisher线性鉴别分析(LDA)在人脸等高维图像识别应用中不可避免地遇到小样本问题。提出一种基于散度差准则的鉴别分析方法。与LDA方法不同的是,该方法利用样本模式的类间散布与类内散布之差而不是它们的比作为鉴别准则,这样,从根本上避免了类内散布矩阵奇异带来的困难。在ORL人脸数据库和AR人脸数据库上的实验结果验证算法的有效性。It is well-known that the applicability of Fisher linear discriminant analysis（LDA）to high-dimensional image recognition tasks such as face recognition inevitably suffers from the so-called ＂small sample size＇（SSS）problem.In this paper,We propose a novel discriminant analysis method to essentially avoid the SSS problem using scatter difference discriminant criterion.Different from LDA,the method adopts the difference of between-class scatter and withln-class scatter as diseriminant criterion rather than the ratio of them.Extensive experiments performed on both ORL face database and AR face database verify the effectiveness of the proposed method.

关 键 词：FISHER线性鉴别分析 最大散度差鉴别分析 人脸识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]