基于二维复判别分析的人脸识别研究  

Study on face recognition based on two-dimensional combined complex discriminant analysis

作　　者：胡晓[1] 俞王新[2] 余群[1] 

机构地区：[1]广州大学电子信息工程系,广东广州510006 [2]上海交通大学生物医学工程系,上海200240

出　　处：《计算机工程与设计》2010年第11期2514-2518,共5页Computer Engineering and Design

基　　金：广东省教育部科技部企业科技特派员行动计划专项基金项目(2009B090600034);广东省科技计划基金项目(2009B060700124);广州市教育科学"十一五"规划基金项目(07B171)

摘　　要：为了提高人脸正确识别率和效率,在行列方向的二维线性判别分析((2D)2LDA)基础之上,提出了一种二维复判别分析(2DCCDA)的人脸识别方法。该方法通过(2D)2LDA并行提取到的行和列特征矩阵,利用复二维鉴别式分析(C2DLDA)将行和列特征融合成复数特征矩阵,从复数特征矩阵中提取出最具分类能力的系数组成特征向量。相比较二维线性判别分析(2DLDA)和(2D)2LDA方法,2DCCDA需要更少的特征系数来表征一幅图像,并且正确识别率也相应提高。To increase the accurate recognition rate and the efficiency of face recognition,on the basis of two-dimensional linear discriminant analysis（2DLDA） along row and column directions（（2D）2LDA）,a face recognition method is proposed,based on two-dimensional combined complex discriminant analysis（2DCCDA）.In this method,firstly row and column feature matrixes are extracted by（2D）2LDA.Secondly,both feature matrixes are integrated into a complex feature matrix using complex version of 2DLDA（C2DLDA）.In the end,some components are picked out from the complex feature matrix to form a feature vector according to their discriminative abilities.Compared with 2DLDA and（2D）2LDA methods,the proposed 2DCCDA can present an image with less feature components and the accurate recognition rate is improved.

关 键 词：人脸识别 主成份分析 线性判别分析 复二维鉴别式分析 二维复判别分析 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]