摘要： 线性鉴别分析(LDA)小样本问题的已有解决方法在构造最优投影子空间时未完整利用LDA 的4 个信息空间,为此,提出一种基于二维主成分分析(2D-PCA)的两级LDA 人脸识别方法。采用减法运算对样本类内散度矩阵和类间散度矩阵的特征值矩阵求逆,以解决小样本问题,并连续应用Fisher 准则和修改后的Fisher 准则连接2 个投影子空间,获取包含LDA 的4 个信息空间的最优投影方向,利用2D-PCA 对输入样本做预处理,以减少计算复杂度。在ORL 和YALE 人脸库上的实验结果表明,该方法虽然训练时间略有增加,但识别率分别为92. 5% 和95. 8% ,优于其他常用LDA 算法。

关键词: 线性鉴别分析, 直接线性鉴别分析, 二维主成分分析, 小样本问题, 人脸识别, 特征提取

Abstract: Aiming at the existing algorithms which do not use the whole four information space of Linear Discriminant Analysis(LDA) in solving the small sample size problem,a two-stage LDA face recognition algorithm based on Two Dimension Principle Component Analyses (2D-PCA ) is proposed. The small sample size problem is solved by a subtraction to estimate the inverse matrix of the eigenvalues matrix of the singular with-class scatter matrix and betweenclass scatter matrix. Thus,the projection subspaces resulting from continuously using the traditional Fisher criterion and a modified Fisher criterion,are concatenated to obtain the optimal projection space including whole four information space of LDA. To reduce the computational complexity,the 2D-PCA is used to preprocess on input samples. The recognize rates of the proposed algorithm on ORL and YALE database are 92. 5% and 95. 8% which are higher than other LDA algorithms despite the slightly increase of training time.

Key words: Linear Discriminant Analysis ( LDA ), Direct LDA ( DLDA ), Two Dimension Principle Component Analysis(2D-PCA), small sample size problem, face recognition, feature extraction

中图分类号: 

• TP18