关键词: 主成分分析, 模块主成分分析, 特征抽取, 人脸识别

Abstract: An improved modular PCA（Principal Component Analysis） method，that is modular PCA method based on independence of feature extraction，is proposed.The original images are divided into sub-images in proposed approach.Then each kind of sub-images at the same position have been disposed by PCA independently，the distance between the corresponding sub-images of the test sample and the train sample can be given.Finally，the distance between the test sample and the train sample can be caculated by adding all these distances between the sub-images together，the nearest distance classification is used to distinguish each face.Experimental results on ORL face database and Yale face database indicate that the improved modular PCA is obviously superior to that of general modular PCA.

Key words: Principal Component Analysis（PCA）, modular principal component analysis, feature extraction, face recognition