2DPCA versus PCA for face recognition

Abstract

Dimensionality reduction methods play an important role in face recognition. Principal component analysis (PCA) and two-dimensional principal component analysis (2DPCA) are two kinds of important methods in this field. Recent research seems like that 2DPCA method is superior to PCA method. To prove if this conclusion is always true, a comprehensive comparison study between PCA and 2DPCA methods was carried out. A novel concept, called column-image difference (CID), was proposed to analyze the difference between PCA and 2DPCA methods in theory. It is found that there exist some restrictive conditions when 2DPCA outperforms PCA. After theoretical analysis, the experiments were conducted on four famous face image databases. The experiment results confirm the validity of theoretical claim.

Keywords

face recognition / dimensionality reduction / 2DPCA method / PCA method / column-image difference (CID)

References

[1] TurkM, PentlandA. Eigenfaces for recognition [J]. Journal of Cognitive Neuroscience, 1991, 3(1): 71-86

[2] MartínezA M, KakA C. PCA versus LDA [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2001, 23(2): 228-233

[3] JungD Y, LeeS M, KimJ H, LeeS H. Fault detection method with PCA and LDA and its application to induction motor [J]. Journal of Central South University of Technology, 2010, 17(6): 1238-1242

[4] HeX, YanS, HuY, NiyogiP, ZhangH J. Face recognition using Laplacianfaces [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, 27(3): 328-340

[5] QiaoL, ChenS, TanX. Sparsity preserving projections with applications to face recognition [J]. Pattern Recognition, 2010, 43(1): 331-341

[6] YangJ, ZhangD, FrangiA F, YangJ Y. Two-dimensional PCA: A new approach to appearance-based face representation and recognition [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2004, 26(1): 131-137

[7] KongH, WangL, TeohE K, LiX, WangJ G. Generalized 2D principal component analysis for face image representation and recognition [J]. Neural Networks, 2005, 18(5): 585-594

[8] YeJ. Generalized low rank approximations of matrices [J]. Machine Learning, 2005, 61(1/2/3): 167-191

[9] LiM, YuanB. 2D-LDA: A statistical linear discriminant analysis for image matrix [J]. Pattern Recognition Letters, 2005, 26(5): 527-532

[10] GuoZ Q, YangJ, JiaoH F. Face recognition method based on two-dimensional locality preserving projections [J]. Computer Engineering, 2011, 37(7): 4-6

[11] LiX, PangY, YuanY. L1-norm-based 2DPCA [J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2010, 40(4): 1170-1175

[12] WangH, WangJ. 2DPCA with L1-norm for simultaneously robust and sparse modeling [J]. Neural Networks, 2013, 46: 190-198

[13] ZhengH, PanL. An improved block 2DPCA face recognition algorithm with L1-norm [J]. Advanced Materials Research, 2012, 1639(457): 1077-1082

[14] ZhangY, QiM X, ShangL. Palmprint recognition based on two-dimensional gabor wavelet transform and two-dimensional principal component analysis [M]. Advanced Intelligent Computing, 2012, Berlin Heidelberg, Springer: 405-411

[15] HeG, TangY, FangB, WangP S. Bionic face recognition using Gabor transformation [J]. International Journal of Pattern Recognition and Artificial Intelligence, 2011, 25(3): 391-402

[16] WangD G, YangZ L, ChangS. SAR image recognition combined bidirectional 2DPCA with PCA [J]. Advanced Materials Research, 2013, 2534(756): 4041-4044

[17] ZhaoZ Q, HuangD S, JiaW. Palmprint recognition with 2DPCA+PCA based on modular neural networks [J]. Neurocomputing, 2007, 71(1): 448-454

[18] WahyuningrumR T, DamayantiF. Efficient kernel-based two-dimensional principal component analysis for smile stages recognition [J]. Telkomnika, 2012, 10(1): 113-118

[19] ChoiY, OzawaS, LeeM. Incremental two-dimensional kernel principal component analysis [J]. Neurocomputing, 2014, 134: 280-288

[20] Da-zhengZ Y F. An algorithm of feature extraction of face based on the weighted variation of 2DPCA [J]. Journal of Electronics & Information Technology, 2011, 33(4): 769-774

[21] RouabhiaC, TebbikhH. Efficient face recognition based on weighted matrix distance metrics and 2DPCA algorithm [J]. Archives of Control Sciences, 2011, 21(2): 207-221

[22] ChenC M, SongJ T, ZhangS Q. Face recognition method based on 2DPCA and compressive sensing [J]. Computer Engineering, 2011, 33(22): 176-178

[23] MengJ, ZhengX. Robust sparse 2D principal component analysis for object recognition [J]. Applied Mathematics & Information Sciences, 2013, 7(6): 2509-2514

[24] ZhangD, ZhouZ H. (2D) 2PCA: Two-directional PCA for efficient face representation and recognition [J]. Neurocomputing, 2005, 69(1): 224-231

[25] MARTINEZ A M, BENAVENTE R. The AR face database [R]. CVC Technical Report # 24, June 1998.

[26] GrahamD B, AllinsonN M. Face recognition: From theory to applications [J]. NATO ASI Series F: computer and Systems Sciences, 1998, 163: 446-456