Email this article A Comparative Study on Vector-based and Matrix-based Linear Discriminant Analysis
Abstract
singularity problem and has low computational costs and has been experimentally reported that 2DLDA outperforms traditional LDA. In this paper, we compare 2DLDA with
LDA in view of the discriminant power and find that 2DLDA as a kind of special LDA has no stronger discriminant power than LDA. So, why 2DLDA outperforms LDA in some cases?
Through theoretical analysis, we find it is mainly because of the difference of stability under nonsingular linear transformation and linear operation power between 2DLDA
and LDA. In experimental parts, the results of experiments give enough proof on our claims and show in some cases the performance of 2DLDA will be possible superior to that of
LDA and in other cases the performance of LDA will be possible superior to that of 2DLDA.
Keywords
References
[1] P.N. Belhumeur, J. Hespanda, D. Kriegeman, “Eigenfaces vs Fisherfaces: Recognition using class specific linear projection,” IEEE Trans. Pattern Anal. Mach. Intell. London, vol. 19, pp. 711-720, August 1997.
doi:10.1109/34.598228
[2] H. Cevikalp, M. Neamtu, M. Wilkes, A. Barkana, “Discriminative common vectors for face recognition,” IEEE Trans. Pattern Anal. Mach. Intell. London, vol. 27, pp. 4-13, September 2005.
doi:10.1109/TPAMI.2005.9
PMid:15628264
[3] L. Chen, H. Liao, M. Ko, J. Lin, G. Yu, “A new LDA based face recognition system which can solve the small sample size problem,” Pattern Recognition. London, vol. 33, pp. 1713-1726, December 2000.
doi:10.1016/S0031-3203(99)00139-9
[4] R. Huang, Q.S. Liu, H.Q. Lu, S.D. Ma, “Solving the small sample size problem of LDA,” in: ICPR., vol. 3, pp. 29-32, 2002.
[5] J. Lu, K.N. Plataniotis, A.N. Venetsanopoulos, “Regularization studies of linear discriminant analysis in small sample size scenarios with application to face recognition,” Pattern Recognition Letters. London, vol. 26, pp. 181-191, January 2005.
doi:10.1016/j.patrec.2004.09.014
[6] D.Q. Dai, P.C. Yuen, “Regularized discriminant analysis and its application to face recognition,” Pattern Recognition. London, vol. 36, pp. 845-847, March 2003.
doi:10.1016/S0031-3203(02)00092-4
[7] W. Zhao, R. Chellappa, P.J. Phillips, “Subspace linear discriminant analysis for face recognition,” Technical Report CAR- TR-914, CS-TR-4009, University of Maryland, College Park, MD.
[8] P. Zhang, J. Peng, N. Riedel, “Discriminant analysis: a least squares approximation view,” in: CVPR, pp. 46-46, 2005.
[9] J. Yang, J.Y. Yang, “From image vector to matrix: a straightforward image projection technique—IMPCA vs. PCA,” Pattern Recognition. vol. 35, pp. 1997-1999, September 2002.
doi:10.1016/S0031-3203(02)00040-7
[10] J. Yang, D. Zhang, A.F. Frangi, J.Y. Yang, “Twodimensional PCA: a new approach to appearance-based face representation and recognition,” IEEE Trans. Pattern Anal. Mach. Intell. vol. 26, pp. 131-137, 2004.
doi:10.1109/TPAMI.2004.1261097
PMid:15382693
[11] M. Li, B. Yuan, “2D-LDA: a novel statistical linear discriminant analysis for image matrix,” Pattern Recognition Letter. vol. 26, pp. 527-532, 2005.
doi:10.1016/j.patrec.2004.09.007
[12] H. Kong, L. Wang, E. Teoh, J. Wang, V. Ronda, “Generalized 2D principal component analysis,” in: IEEE Conference on IJCNN. Canada, vol. 1, pp. 108-113, 2005.
[13] H. Xiong, M.N.S Swamy, M.O. Ahmad, “Two-dimensional FLD for face recognition,” Pattern Recognition. vol. 38, pp. 1121-1124, July 2005.
doi:10.1016/j.patcog.2004.12.003
[14] J. Ye, R. Janardan, Q. Li, “Two-dimensional linear discriminant analysis,” in: NIPS. 2004.
[15] S. Noushatha, Hemantha, G. Kumar, P. Shivakumara, “(2D)2 LDA: an efficient approach for face recognition,” Pattern Recognition. vol. 39, pp. 1396-1400, July 2006.
doi:10.1016/j.patcog.2006.01.018
[16] S.B. Chen, H.F. Zhao, M. Kong, B. Luo, “2D-LPP: A twodimensional extension of locality preserving projections,” Neurocomputing. vol. 70, pp. 912-921, January 2007.
[17] X. Pan, Q.Q. Ruan, “Palmprint recognition with improved two-dimensional locality preserving projections,” Image and Vision Computing. vol. 26, pp. 1261-1268, September 2008.
doi:10.1016/j.imavis.2008.03.001
[18] W.S. Zheng, J.H. Lai, S.Z. Li, “1D-LDA vs. 2DLDA: When is vector-based linear discriminant analysis better than matrix-based?,” Pattern Recognition. vol. 41, pp. 2156-2172, July 2008.
doi:10.1016/j.patcog.2007.11.025
[19] Z.Z. Liang, Y.F. Li, Shi P F, “A note on two-dimensional linear discriminant analysis,” Pattern Recognition Letters, vol. 29, pp. 2122-2128, December 2008.
doi:10.1016/j.patrec.2008.07.009
[20] G.S. Wang, X. Wu, Z. Jia, “Matrix Inequality”, www.sciencep.com, in press.
[21] ORL, The ORL face database at the AT&T (Olivetti) research laboratory, 1992.
[22] J.P. Ye, R. Janardan, Q. Li, Park H. “Feature reduction via generalized uncorrelated linear discriminant analysis,” IEEE Trans. Knowledge and Data Engineering. vol. 18, pp. 1312-1322, 2006.
doi:10.1109/TKDE.2006.160
Full Text: PDF