基于MW（2D）^2PCA的单训练样本人脸识别  

MW（2D）^2PCA Based Face Recognition with Single Training Sample

作　　者：李欣[1,2] 王科俊[2] 贲晛烨 

机构地区：[1]哈尔滨工程大学工程训练中心,哈尔滨150001 [2]哈尔滨工程大学自动化学院,哈尔滨150001

出　　处：《模式识别与人工智能》2010年第1期77-83,共7页Pattern Recognition and Artificial Intelligence

基　　金：国家863计划资助项目(No.2006AA04Z248)

摘　　要：传统的人脸识别方法在单训练样本条件下性能会急剧下降,因此,研究出适合于单样本情况下的识别算法是人脸识别问题面临的巨大挑战.针对两个方向的二维主成分分析((2D)2PCA)算法进行改进,文中提出将加权和分块与(2D)2PCA相结合的方法称为分块加权(2D)2PCA,以便更有效地提取人脸的局部特征.同时把模糊理论引入分类决策,应用于单训练样本人脸识别问题.在ORL人脸库以及部分CAS-PEAL人脸库中的实验结果表明,文中方法能取得较好的识别效果.Traditional methods get low recognition accuracy in the condition of only one training sample, therefore, it is a great challenge for face recognition. In this paper, a two-directional two-dimensional principal component analysis（（2D）^2PCA） is developed to solve this problem. An improved arithmetic combining weight and block called modular weighted （2D）^2PCA is proposed for efficient local feature extraction. Besides, the fuzzy theory is introduced to classify the single sample face recognition. Experimental results on ORL and a subset of CAS-PEAL face databases show that the presented method achieves a high recognition accuracy.

关 键 词：单样本人脸识别 局部特征提取 主成分分析(PCA) 两个方向的二维主成分分析((2D)2PCA) 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]