Skip to main content
Log in

Generalized two-dimensional PCA based on \(\ell _{2,p}\)-norm minimization

  • Original Article
  • Published:
International Journal of Machine Learning and Cybernetics Aims and scope Submit manuscript

Abstract

To exploit the information from two-dimensional structured data, two-dimensional principal component analysis (2-DPCA) has been widely used for dimensionality reduction and feature extraction. However, 2-DPCA is sensitive to outliers which are common in real applications. Therefore, many robust 2-DPCA methods have been proposed to improve the robustness of 2-DPCA. But existing robust 2-DPCAs have several weaknesses. First, these methods cannot be robust enough to outliers. Second, to center a sample set mixed with outliers using the L2-norm distance is usually biased. Third, most methods do not preserve the nice property of 2-DPCA (rotational invariance), which is important for learning algorithm. To alleviate these issues, we present a generalized robust 2-DPCA, which is named as 2-DPCA with \(\ell _{2,p}\)-norm minimization (\(\ell _{2,p}\)-2-DPCA), for image representation and recognition. In \(\ell _{2,p}\)-2-DPCA, \(\ell _{2,p}\)-norm is employed as the distance metric to measure the reconstruction error, which can alleviate the effect of outliers. Therefore, the proposed method is robust to outliers and preserves the desirable property of 2-DPCA which is invariant to rotational and well characterizes the geometric structure of samples. Moreover, most existing robust PCA methods estimate sample mean from database with outliers by averaging, which is usually biased. Sample mean are treated as an unknown variable to remedy the bias of computing sample mean in \(\ell _{2,p}\)-2-DPCA. To solve \(\ell _{2,p}\)-2-DPCA, we propose an iterative algorithm, which has a closed-form solution in each iteration. Experimental results on several benchmark databases demonstrate the effectiveness and advantages of our method.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6

Similar content being viewed by others

References

  1. Chang X, Nie F, Wang S, Yang Y, Zhou X, Zhang C (2016) Compound rank-\(k\) projections for bilinear analysis. IEEE Trans Neural Netw Learn Syst 27(7):1502–1513

    MathSciNet  Google Scholar 

  2. Parsons L, Haque E, Liu H (2004) Subspace clustering for high dimensional data: a review. ACM Sigkdd Explor Newsl 6(1):90–105

    Google Scholar 

  3. Chang X, Nie F, Yang Y, Zhang C, Huang H (2016) Convex sparse PCA for unsupervised feature learning. ACM Trans Knowl Discov Data (TKDD) 11(1):3

    Google Scholar 

  4. Shi X, Xing F, Guo Z, Su H, Liu F, Yang L (2019) Structured orthogonal matching pursuit for feature selection. Neurocomputing 349:164–172

    Google Scholar 

  5. Wold S, Esbensen K, Geladi P (1987) Principal component analysis. Chemom Intell Lab Syst 2(1–3):37–52

    Google Scholar 

  6. Belhumeur PN, Hespanha JP, Kriegman DJ (1997) Eigenfaces vs. fisherfaces: recognition using class specific linear projection. Technical report. Yale University New Haven United States

  7. He X, Cai D, Yan S, Zhang HJ (2005) Neighborhood preserving embedding. In: Tenth IEEE international conference on computer vision (ICCV’05) volume 1, vol 2. IEEE, pp 1208–1213

  8. He X, Niyogi P (2004) Locality preserving projections. In: Advances in neural information processing systems, pp 153–160

  9. Belkin M, Niyogi P (2003) Laplacian eigenmaps for dimensionality reduction and data representation. Neural Comput 15(6):1373–1396

    MATH  Google Scholar 

  10. Roweis ST, Saul LK (2000) Nonlinear dimensionality reduction by locally linear embedding. Science 290(5500):2323–2326

    Google Scholar 

  11. Liu H, Lai Z, Chen Y (2019) Joint sparse neighborhood preserving embedding. J Phys Conf Ser 1176:032023

    Google Scholar 

  12. Liu H, Lai Z, Chen Y (2017) Joint sparse locality preserving projections. In: International conference on smart computing and communication. Springer, pp 125–133

  13. Lu Y, Yuan C, Lai Z, Li X, Wong WK, Zhang D (2017) Nuclear norm-based 2dlpp for image classification. IEEE Trans Multimed 19(11):2391–2403

    Google Scholar 

  14. Ding C, Zhou D, He X, Zha H (2006) R 1-PCA: rotational invariant l 1-norm principal component analysis for robust subspace factorization. In: Proceedings of the 23rd international conference on Machine learning. ACM, pp 281–288

  15. Kwak N (2008) Principal component analysis based on l1-norm maximization. IEEE Trans Pattern Anal Mach Intell 30(9):1672–1680

    Google Scholar 

  16. Shi X, Guo Z, Nie F, Yang L, You J, Tao D (2015) Two-dimensional whitening reconstruction for enhancing robustness of principal component analysis. IEEE Trans Pattern Anal Mach Intell 38(10):2130–2136

    Google Scholar 

  17. Candès EJ, Li X, Ma Y, Wright J (2011) Robust principal component analysis? J ACM (JACM) 58(3):11

    MathSciNet  MATH  Google Scholar 

  18. Liu JX, Xu Y, Zheng CH, Kong H, Lai ZH (2015) Rpca-based tumor classification using gene expression data. IEEE/ACM Trans Comput Biol Bioinform (TCBB) 12(4):964–970

    Google Scholar 

  19. Shi X, Nie F, Lai Z, Guo Z (2018) Robust principal component analysis via optimal mean by joint \(\ell _{2,1}\) and schatten p-norms minimization. Neurocomputing 283:205–213

    Google Scholar 

  20. Wang Q, Gao Q, Gao X, Nie F (2016) \(\ell\)2, p-norm based pca for image recognition. IEEE Trans Image Process 27:1

    MathSciNet  Google Scholar 

  21. Baccini A, Besse P, Falguerolles A (1996) A l1-norm pca and a heuristic approach. Ordinal Symb Data Anal 1(1):359–368

    MATH  Google Scholar 

  22. Nie F, Huang H, Ding C, Luo D, Wang H (2011) Robust principal component analysis with non-greedy l1-norm maximization. In: IJCAI Proceedings-international joint conference on artificial intelligence, vol 22, p 1433

  23. Markopoulos PP, Kundu S, Chamadia S, Pados DA (2017) Efficient l1-norm principal-component analysis via bit flipping. IEEE Trans Signal Process 65(16):4252–4264

    MathSciNet  MATH  Google Scholar 

  24. Park YW, Klabjan D (2016) Iteratively reweighted least squares algorithms for l1-norm principal component analysis. In: 2016 IEEE 16th international conference on data mining (ICDM). IEEE, pp 430–438

  25. Brooks JP, Dulá J, Boone EL (2013) A pure l1-norm principal component analysis. Comput Stat Data Anal 61:83–98

    MATH  Google Scholar 

  26. Lu GF, Zou J, Wang Y, Wang Z (2016) L1-norm-based principal component analysis with adaptive regularization. Pattern Recognit 60:901–907

    MATH  Google Scholar 

  27. Tsagkarakis N, Markopoulos PP, Sklivanitis G, Pados DA (2018) L1-norm principal-component analysis of complex data. IEEE Trans Signal Process 66(12):3256–3267

    MathSciNet  MATH  Google Scholar 

  28. Pang Y, Li X, Yuan Y (2010) Robust tensor analysis with l1-norm. IEEE Trans Circuits Syst Video Technol 20(2):172–178

    Google Scholar 

  29. Yi S, Lai Z, He Z, Ym Cheung, Liu Y (2017) Joint sparse principal component analysis. Pattern Recognit 61:524–536

    MATH  Google Scholar 

  30. Kwak N (2014) Principal component analysis by \(l\_\)\(p\)-norm maximization. IEEE Trans Cybern 44(5):594–609

    Google Scholar 

  31. Shi X, Guo Z, Xing F, Cai J, Yang L (2018) Self-learning for face clustering. Pattern Recognit 79:279–289

    Google Scholar 

  32. Li X, Pang Y, Yuan Y (2010) L1-norm-based 2dpca. IEEE Trans Syst Man Cybern Part B (Cybern) 40(4):1170–1175

    Google Scholar 

  33. Yang J, Zhang D, Frangi AF, Jy Yang (2004) Two-dimensional pca: a new approach to appearance-based face representation and recognition. IEEE Trans Pattern Anal Mach Intell 26(1):131–137

    Google Scholar 

  34. Kong H, Wang L, Teoh EK, Li X, Wang JG, Venkateswarlu R (2005) Generalized 2d principal component analysis for face image representation and recognition. Neural Netw 18(5–6):585–594

    Google Scholar 

  35. Gu Z, Shao M, Li L, Fu Y (2012) Discriminative metric: Schatten norm vs. vector norm. In: 2012 21st international conference on pattern recognition (ICPR). IEEE, pp 1213–1216

  36. Zhang F, Yang J, Qian J, Xu Y (2015) Nuclear norm-based 2-dpca for extracting features from images. IEEE Trans Neural Netw Learn Syst 26(10):2247–2260

    MathSciNet  Google Scholar 

  37. Chen Y, Lai Z, Zhang Y (2016) Sparse nuclear norm two dimensional principal component analysis. In: Chinese conference on biometric recognition. Springer, pp 547–555

  38. Chen Y, Lai Z, Wen J, Gao C (2018) Nuclear norm based two-dimensional sparse principal component analysis. Int J Wavelets Multiresolution Inf Process 16(02):1840002

    MathSciNet  MATH  Google Scholar 

  39. Mi JX, Zhang YN, Lai Z, Li W, Zhou L, Zhong F (2019) Principal component analysis based on nuclear norm minimization. Neural Netw 118:1–16

    MATH  Google Scholar 

  40. Wang H, Wang J (2013) 2dpca with l1-norm for simultaneously robust and sparse modelling. Neural Netw 46:190–198

    MATH  Google Scholar 

  41. Lai Z, Xu Y, Chen Q, Yang J, Zhang D (2014) Multilinear sparse principal component analysis. IEEE Trans Neural Netw Learn Syst 25(10):1942–1950

    Google Scholar 

  42. Lai Z, Xu Y, Yang J, Shen L, Zhang D (2016) Rotational invariant dimensionality reduction algorithms. IEEE Trans Cybern 47(11):3733–3746

    Google Scholar 

  43. Li Q, Shi X, Zhou L, Bao Z, Guo Z (2017) Active learning via local structure reconstruction. Pattern Recognit Lett 92:81–88

    Google Scholar 

  44. Shi X, Yang Y, Guo Z, Lai Z (2014) Face recognition by sparse discriminant analysis via joint l2, 1-norm minimization. Pattern Recognit 47(7):2447–2453

    Google Scholar 

  45. Wen J, Lai Z, Wong WK, Cui J, Wan M (2014) Optimal feature selection for robust classification via l2, 1-norms regularization. In: 2014 22nd international conference on pattern recognition. IEEE, pp 517–521

  46. Xu X, Lai Z, Chen Y, Kong H (2018) Robust discriminative principal component analysis. In: Chinese conference on biometric recognition. Springer, pp 231–238

  47. Shi X, Guo Z, Lai Z, Yang Y, Bao Z, Zhang D (2015) A framework of joint graph embedding and sparse regression for dimensionality reduction. IEEE Trans Image Process 24(4):1341–1355

    MathSciNet  MATH  Google Scholar 

  48. Wang Q, Gao Q (2017) Two-dimensional pca with f-norm minimization. In: AAAI, pp 2718–2724

  49. Gao Q, Xu S, Chen F, Ding C, Gao X, Li Y (2018) \(r_1\)-2-dpca and face recognition. IEEE Trans Cybern 49(4):1212–1223

    Google Scholar 

  50. Chartrand R, Yin W (2008) Iteratively reweighted algorithms for compressive sensing. In: 2008 IEEE international conference on acoustics, speech and signal processing. IEEE, pp 3869–3872

  51. Lee KC, Ho J, Kriegman DJ (2005) Acquiring linear subspaces for face recognition under variable lighting. IEEE Trans Pattern Anal Mach Intell 5:684–698

    Google Scholar 

  52. Sim T, Baker S, Bsat M (2002) The cmu pose, illumination, and expression (pie) database. In: Proceedings of the fifth IEEE international conference on automatic face and gesture recognition. IEEE, pp 53–58

  53. Martinez AM (1998) The ar face database. CVC Technical report 24

  54. Nene SA, Nayar SK, Murase H et al (1996) Columbia object image library (coil-20)

  55. Nefian AV (2013) Georgia tech face database

  56. Keller JM, Gray MR, Givens JA (1985) A fuzzy k-nearest neighbor algorithm. IEEE Trans Syst Man Cybern 4:580–585

    Google Scholar 

Download references

Acknowledgements

This work was supported by Natural Science Foundation of Chongqing (under Grant Nos. cstc2018jcyjAX0532 and cstc2016jcyjA0407) and National Natural Science Foundation of China 61906024.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Jian-Xun Mi.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Mi, JX., Zhang, YN., Li, Y. et al. Generalized two-dimensional PCA based on \(\ell _{2,p}\)-norm minimization. Int. J. Mach. Learn. & Cyber. 11, 2421–2438 (2020). https://doi.org/10.1007/s13042-020-01127-1

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s13042-020-01127-1

Keywords

Navigation