Linear Discriminant Analysis(LDA) is one of the most popular linear classification techniques for feature extraction,but when dealing with face recognition, it will meet two problems:small sample size and rank limitation. In order to solve these two problems,this paper presents a modified LDA based on linear combination of k-order matrices-MLDA.MLDA redefines within-class scatter matrix Sw in order to make the traditional Fisher criterion get much more suitable to other situations. Experiments on different face databases verify the effectiveness of MLDA.%线性判别分析(LDA)是一种普遍用于特征提取的线性分类方法.但将 LDA 直接用于人脸识别会遇到小样本问题和秩限制问题.为了解决以上问题,提出一种基于多阶矩阵组合的 LDA 算法--MLDA.该算法重新定义了传统 LDA 中的类内离散度矩阵S,使传统 Fisher 准则具有更好的健壮性和适应性.若干人脸数据库上的比较实验证明了 MLDA 有效性.
展开▼