Title :
Two-dimensional neighborhood preserving discriminant analysis for face recognition
Author :
Lu, Guanming ; Zuo, Jiakuo
Author_Institution :
Coll. of Telecommun. & Inf. Eng., Nanjing Univ. of Posts & Telecommun., Nanjing, China
Abstract :
In this paper, we propose an innovative feature extraction algorithm named two-dimensional neighborhood preserving discriminant analysis (2DNPDA), which directly extracts feature from image matrix. The proposed algorithm considers both the neighborhood structure of the samples and the discriminant information of different classes. Experimental results on ORL and Yale face databases show that 2DNPDA can attain better recognition rate than PCA, LDA, MMC, 2DPCA, 2DLDA and LPP.
Keywords :
face recognition; feature extraction; 2DLDA; 2DPCA; ORL face databases; PCA; Yale face databases; face recognition; image matrix; innovative feature extraction algorithm; linear discriminant analysis; locality preserving projection; maximum margin criterion; two-dimensional neighborhood preserving discriminant analysis; Databases; Face; Face recognition; Feature extraction; Image reconstruction; Principal component analysis; Training; Face recognition; Feature extraction; Linear discriminant analysis (LDA); Locality preserving projection (LPP); Maximum margin criterion (MMC); Two-dimensional neighborhood preserving discriminant analysis (2DNPDA);
Conference_Titel :
Natural Computation (ICNC), 2010 Sixth International Conference on
Conference_Location :
Yantai, Shandong
Print_ISBN :
978-1-4244-5958-2
DOI :
10.1109/ICNC.2010.5582718
