Face Recognition Based on 2DPCA under MMSE Rule and RBF Neural Network

Author

Luo, Renze ; Ran, Ruisheng ; Wang, Ruyan

Author_Institution

Electron. Eng. Dept., UESTC, Zhongshan

Volume

3

fYear

2009

fDate

7-8 March 2009

Firstpage

804

Lastpage

807

Abstract

In this paper, based on the rule of minimal mean square error(MMSE), a method for face image feature extraction-2DPCA, is presented. Meanwhile, in theory, it is firstly proved that face feature extraction of 2DPCA ensures the energy lost is minimal and this method is reasonable; Then, RBF neural network is used to classify the test image. Experiment results show that, for the given threshold, the recognition rate of 2DPCA is about 95% and compared with PCA, much less running time is required for 2DPCA.

Keywords

face recognition; feature extraction; image classification; mean square error methods; neural nets; principal component analysis; radial basis function networks; RBF neural network; feature extraction; feature extraction-2DPCA; image classification; minimal mean square error method; principal component analysis; radial basis function network; Computer science education; Educational technology; Electronic mail; Face recognition; Feature extraction; Mean square error methods; Neural networks; Optical scattering; Principal component analysis; Testing; 2DPCA; face recognition; the rule of MMSE;

fLanguage

English

Publisher

ieee

Conference_Titel

Education Technology and Computer Science, 2009. ETCS '09. First International Workshop on

Conference_Location

Wuhan, Hubei

Print_ISBN

978-1-4244-3581-4

Type

conf

DOI

10.1109/ETCS.2009.715

Filename

4959432