二维主成分分析方法的推广及其在人脸识别中的应用

被引:20
作者
陈伏兵
陈秀宏
高秀梅
杨静宇
机构
[1] 南京理工大学计算机科学系
关键词
线性鉴别分析; 特征抽取; 分块二维主成分分析; 特征矩阵; 人脸识别;
D O I
暂无
中图分类号
TP391.41 [];
学科分类号
080203 ;
摘要
提出了分块二维主成分分析(分块2DPCA)的人脸识别方法。分块2DPCA方法先对图像矩阵进行分块,对分块得到的子图像矩阵直接进行鉴别分析。其特点是:能方便地降低鉴别特征的维数;可以完全避免使用矩阵的奇异值分解,特征抽取方便;与2DPCA方法相比,使用低维的鉴别特征矩阵,而达到较高(至少是不低)的正确识别率。此外,2DPCA是分块2DPCA的特例。在ORL和NUST603人脸库上的试验结果表明,所提出的方法在识别性能上优于2DPCA方法。
引用
收藏
页码:1767 / 1770
页数:4
相关论文
共 9 条
  • [1] An efficient algorithm for FoleySammon optimal set of discrim inant vectors by algebraic method. LIU K,CHENG YQ,YANG JY. International Journal of Pattern Recognition and Artificial Intelligence . 1992
  • [2] A lgebraic feature extraction for image recognition based on an optimal discrim inant criterion. LIU K,CHENG YQ,YANG JY. Pattern Recognition . 1993
  • [3] Two-dimensional PCA: A new approach to appearance-based face representation and recognition. YANG J,ZHANG D,YANG JY. IEEE Transactions Pattern Anal. Machine Intell . 2004
  • [4] Why can LDA be performed in PCA transformed space?. YANG J,YANG JY. Pattern Recognition . 2003
  • [5] Looking at people: Sensing for ubiquitous and wearable computing. PENTLAND A. IEEE Transactions on Pattern AnalMachine Intell . 2000
  • [6] A new LDA-based face recognition system which can solve the small sample size problem. CHEN LF,YUAN H,LIAO M,et al. Pattern Recognition . 2000
  • [7] Face Recognition based on uncorrelated discrim inant transformation. JIN Z,YANG JY,HU ZS,et al. Pattern Recognition . 2001
  • [8] Optimal discrim inant plane for a small number of samples and design method of classifier on the plane. HONG ZQ,YANG JY. Pattern Recognition . 1991
  • [9] E igenfaces vs Fisherfaces: Recognition using class specific linearprojection. BELHUMEUR PN,HESPANHA JP,KR IENGMAN DJ. IEEE Transactions on Pattern AnalMachine Intell . 1997