Locality preserving projection (LPP) is one of the most representative linear manifold learning methods and well exploits intrinsic structure of data. However, the performance of LPP remarkably degenerate in the presence of outliers. To alleviate this problem, the authors propose a robust LPP, namely LPP-L21. LPP-L21 employs L2-norm as the distance metric in spatial dimension of data and L1-norm as the distance metric over different data points. Moreover, the authors employ L1-norm to construct similarity graph, this helps to improve robustness of algorithm. Accordingly, the authors present an efficient iterative algorithm to solve LPP-L21. The authors’ proposed method not only well suppresses outliers but also retains LPP's some nice properties. Experimental results on several image data sets show its advantages.

References

1. 1)
  - 32. Ding, M., Jia, W., Yao, M.: ‘A projection algorithm with local preservation based on L2 norm’, J. Xi'an Jiaotong Univ., 2016, 50, (2), pp. 33–37.
2. 2)
  - 18. Markopoulos, P.P., Karystinos, G.N., Pados, D.A.: ‘Optimal algorithms for-subspace signal processing’, IEEE Trans. Signal Process., 2014, 62, (19), pp. 5046–5058.
3. 3)
  - 29. Gui, J., Tao, D., Sun, Z., et al: ‘Group sparse multi-view patch alignment framework with view consistency for image classification’, IEEE Trans. Image Process., 2014, 23, (7), pp. 3126–3137.
4. 4)
  - 2. He, X., Yan, S., Hu, Y., et al: ‘Face recognition using Laplacianfaces’, IEEE Trans. Pattern Anal. Mach. Intell., 2005, 27, (3), pp. 328–340.
5. 5)
  - 37. Martinez, A.M.: ‘The AR face database’. CVC Technical Report, 24, 1998.
6. 6)
  - 4. He, X., Cai, D., Yan, S., et al: ‘Neighborhood preserving embedding’. Proc. of IEEE Int. Conf. on Computer Vision, Beijing, China, 2005, vol. 1, no. 2, pp. 1208–1213.
7. 7)
  - 3. He, X., Yan, S., Hu, Y., et al: ‘Learning a locality preserving subspace for visual recognition’. Proc. of IEEE Int. Conf. on Computer Vision, Madison, USA, 2003, pp. 385–392.
8. 8)
  - 7. Lai, Z., Wong, W.K., Xu, Y., et al: ‘Approximate orthogonal sparse embedding for dimensionality reduction’, IEEE Trans. Neural Netw. Learn. Syst., 2017, 27, (4), pp. 723–735.
9. 9)
  - 39. Lazebnik, S., Schmid, C., Ponce, J.: ‘Beyond bags of features: spatial pyramid matching for recognizing natural scene categories’. Proc. IEEE Computer Vision and Pattern Recognition, New York, USA, 2006, pp. 2169–2178.
10. 10)
  - 34. Liao, S., Li, J., Liu, Y., et al: ‘Robust formulation for PCA: avoiding mean calculation with L2p-norm maximization’. AAAI, New Orleans, USA, 2018.
11. 11)
  - 16. Kwak, N.: ‘Principal component analysis based on L1-norm maximization’, IEEE Trans. Pattern Anal. Mach. Intell., 2008, 30, (9), pp. 1672–1680.
12. 12)
  - 14. Wang, R., Nie, F., Yang, X., et al: ‘Robust 2dpca with non-greedy l1-norm maximization for image analysis’, IEEE Trans. Cybern., 2015, 45, (5), pp. 1108–1112.
13. 13)
  - 31. Wong, W., Lai, Z., Xu, Y., et al: ‘Joint tensor feature analysis for visual object recognition’, IEEE Trans. Cybern., 2015, 45, (11), pp. 2425–2436.
14. 14)
  - 38. Sim, T., Bakerand, S., Bsat, M.: ‘The CMU pose, illumination, and expression (PIE) database’. Proc. IEEE Int. Conf. Automatic Face Gesture Recognition, Washington D.C, USA, 2002, pp. 46–51.
15. 15)
  - 12. Gao, Q., Ma, L., Liu, Y., et al: ‘Angle 2DPCA: a new formulation for 2DPCA’, IEEE Trans. Cybern., 2017, 49, pp. 1–7.
16. 16)
  - 35. Wang, H., Yan, S., Xu, D., et al: ‘Trace ratio vs. ratio trace for dimensionality reduction’. Proc. IEEE Int. Conf. Computer Vision and Pattern Recognition, Minneapolis, USA, June 2007, pp. 1–8.
17. 17)
  - 19. Wang, Q., Gao, Q.: ‘2DPCA with F-norm minimization’. AAAI, San Francisco, USA, 2017, pp. 2718–2724.
18. 18)
  - 28. Nie, F., Huang, H., Cai, X., et al: ‘Efficient and robust feature selection via joint L2,1-norms minimization’. Advances in Neural Information Processing Systems, British Columbia, Canada, 2010, pp. 1813–1821.
19. 19)
  - 17. Nie, F., Wang, H., Ding, C.H., et al: ‘Robust principal component analysis with non-greedy l1-norm maximization’. Proc. Int. Joint Conf. on Artificial Intelligence, Barcelona, Spain, 2011, vol. 22, no. 1, pp. 1433–1438.
20. 20)
  - 27. Wang, Q., Gao, Q., Gao, X., et al: ‘L2,p-norm based PCA for image recognition’, IEEE Trans. Image Process., 2017, 27, pp. 1336–1346.
21. 21)
  - 33. Luo, M., Nie, F., Chang, X., et al: ‘Avoiding optimal mean robust PCA/2DPCA with non-greedy l1-norm maximization’. Int. Joint Conf. on Artificial Intelligence, Melbourne, Australia, 2017, pp. 1802–1808.
22. 22)
  - 11. Lai, Z., Mo, D., Wong, W., et al: ‘Robust discriminant regression for feature extraction’, IEEE Trans. Cybern., 2018, 48, pp. 1–13.
23. 23)
  - 22. Liu, Y., Gao, Q., Miao, S., et al: ‘A non-greedy algorithm for L1-LDA’, IEEE Trans. Image Process., 2017, 26, (2), pp. 684–695.
24. 24)
  - 23. Wang, Q., Gao, Q., Xie, D., et al: ‘Robust DLPP with non-greedy ℓ₁-norm minimization and maximization’, IEEE Trans. Neural Netw. Learn. Syst., 2017, 29, pp. 738–743.
25. 25)
  - 24. Pang, Y., Yuan, Y.: ‘Outlier-resisting graph embedding’, Neurocomputing, 2010, 73, (4), pp. 968–974.
26. 26)
  - 15. Ke, Q., Kanade, T.: ‘Robust L1 norm factorization in the presence of outliers and missing data by alternative convex programming’. Proc. of the IEEE Conf. on Computer Vision and Pattern Recognition, San Diego, USA, 2005, pp. 592–599.
27. 27)
  - 13. Wang, H., Wang, J.: ‘2DPCA with L1-norm for simultaneously robust and sparse modeling’, Neural Netw., 2013, 46, (10), pp. 190–198.
28. 28)
  - 6. Gao, Q., Liu, J., Zhang, H., et al: ‘Enhanced fisher discriminant criterion for image recognition’, Pattern Recognit., 2012, 45, (10), pp. 3717–3724.
29. 29)
  - 30. Shi, X., Yang, Y., Guo, Z., et al: ‘Face recognition by sparse discriminant analysis via joint L2,1-norm minimization’, Pattern Recognit., 2014, 47, pp. 2447–2453.
30. 30)
  - 20. Lu, G.F., Zou, J., Wang, Y.: ‘L1-norm and maximum margin criterion based discriminant locality preserving projections via trace lasso’, Pattern Recognit., 2016, 55, pp. 207–214.
31. 31)
  - 8. Yan, S., Xu, D., Zhang, H., et al: ‘Graph embedding and extensions: a general framework for dimensionality reduction’, IEEE Trans. Pattern Anal. Mach. Intell., 2007, 29, (1), pp. 40–51.
32. 32)
  - 1. Turk, M.A., Pentland, A.P.: ‘Face recognition using eigenfaces’. 1991. IEEE Conf. on Computer Vision and Pattern Recognition, Proc. of Computer Vision, Hawaii, USA, 1991, pp. 586–591..
33. 33)
  - 5. He, X., Cai, D., Han, J.: ‘Learning a maximum margin subspace for image retrieval’, IEEE Trans. Knowl. Data Eng., 2008, 20, (2), pp. 189–201.
34. 34)
  - 10. Lin, G., Tang, N., Wang, H.: ‘Locally principal component analysis based on L1-norm maximisation’, IET Image Process., 2015, 9, (2), pp. 91–96.
35. 35)
  - 9. Gao, Q., Gao, F., Zhang, H., et al: ‘Two-dimensional maximum local variation based on image Euclidean distance for face recognition’, IEEE Trans. Image Process., 2013, 22, (10), pp. 3807–3817.
36. 36)
  - 21. Zhong, F., Zhang, J.: ‘Linear discriminant analysis based on L1-norm maximization’, IEEE Trans. Image Process., 2013, 22, (8), pp. 3018–3027.
37. 37)
  - 26. Lai, Z., Xu, Y., Yang, J., et al: ‘Rotational invariant dimensionality reduction algorithms’, IEEE Trans. Cybern., 2017, 47, (11), pp. 3733–3746.
38. 38)
  - 25. Ding, C., Zhou, D., He, X., et al: ‘R1-PCA: rotational invariant L1- norm principal component analysis for robust subspace factorization’. Proc. of the 23rd Int. Conf. on Machine Learning, Pittsburgh, USA, 2006, pp. 281–288.
39. 39)
  - 36. Georghiades, A.S., Belhumeurand, P.N., Kriegman, D.: ‘From few to many: illumination cone models for face recognition under variable lighting and pose’, IEEE Trans. Pattern Anal. Mach. Intell., 2001, 23, (6), pp. 643–660.

Dimensionality reduction by LPP-L21

References

Related content