Since sparse representation (SR) was first introduced into robust face recognition, the argument has lasted for several years about whether sparsity can improve robust face recognition or not. Some work argued that the robust sparse representation (RSR) model has a similar recognition rate as non-sparse solution, while it needs a much higher computational cost due to the larger feature dimensionality in the pixel space. In this study, the authors reveal that the standard RSR model, which expands the dictionary with the identity matrix to reconstruct corruption or occlusion in face images, is essentially a non-sparse solution with a relatively large residual. The reason why the RSR model underperforms may be its inappropriately expanded bases rather than the sparsity itself. Thereby, this study proposes to design a dictionary with an expanded noise bases set which can precisely reconstructs any corruption or occlusion in face images in a subspace. Experimental results show that the algorithm can greatly improve recognition rates for robust face recognition. In addition, the algorithm can be simply performed in a subspace with a small feature dimensionality, thus efficient enough for real systems. This study makes us come to the conclusion that solving the approximation problem in raw pixel space is not necessary for robust face recognition, while solving in a subspace with a much smaller feature dimensionality is enough when the dictionary is well expanded. Finally, this study also confirms that the sparsity plays an important role in SR based classification.

References

1. 1)
  - 24. Davis, G., Mallat, S., Avellaneda, M.: ‘Adaptive greedy approximations’, Constructive Approx., 2004, 13, (1), pp. 57–98 (doi: 10.1007/BF02678430).
2. 2)
  - 18. Zheng, C.-H., Zhang, L., Ng, V.T.-Y., Shiu, S.C.-K., Huang, D.S.: ‘Metasample-based sparse representation for tumor classification’, IEEE/ACM Trans. Comput. Biol. Bioinf., 2011, 8, (5), pp. 1273–1282 (doi: 10.1109/TCBB.2011.20).
3. 3)
  - 21. Zhang, S., Yao, H., Sun, X., et al: ‘Action recognition based on overcomplete independent component analysis’, Inf. Sci., 2014, 281, pp. 635–647 (doi: 10.1016/j.ins.2013.12.052).
4. 4)
  - 36. Wright, J., Yang, A., Ganesh, A., Sastry, S., Ma, Y.: ‘Robust face recognition via sparse representation’, IEEE Trans. Pattern Anal. Mach. Intell., 2009, 31, (2), pp. 210–227 (doi: 10.1109/TPAMI.2008.79).
5. 5)
  - 29. Wu, Y.-C.: ‘A sparse ℓ2-regularized support vector machines for efficient natural language learning’, Knowl. Inf. Syst., 2014, 39, (2), pp. 305–328 (doi: 10.1007/s10115-013-0615-0).
6. 6)
  - 4. Luo, D., Ding, C., Huang, H.: ‘Toward structural sparsity: an explicit ℓ2/ℓ0 approach’, Knowl. Inf. Syst., 2013, 36, (2), pp. 411–438 (doi: 10.1007/s10115-012-0545-2).
7. 7)
  - 28. Belhumeur, P., Hespanha, J., Kriegman, D.: ‘Eigenfaces vs. Fisherfaces: recognition using class specific linear projection’, IEEE Trans. Pattern Anal. Mach. Intell., 1997, 19, pp. 711–720 (doi: 10.1109/34.598228).
8. 8)
  - 27. Deng, W., Hu, J., Guo, J.: ‘In defense of sparsity based face recognition’. CVPR, 2013, pp. 399–406.
9. 9)
  - 8. Zhang, L., Yang, M., Feng, X.: ‘Sparse representation or collaborative representation: Which helps face recognition?’. ICCV, 2011, pp. 471–478.
10. 10)
  - 15. Duda, R., Hart, P., Stork, D.: ‘Pattern classification’ (John Wiley & Sons, 2nd edn.).
11. 11)
  - 11. Fiqueiredo, M.A.T., Nowak, R.D., Wright, S.J.: ‘Gradient projection for sparse reconstruction: Application to compressed sensing and other inverse problems’, IEEE J. Sel. Topics Signal Process., 2007, l, (4), pp. 586–598 (doi: 10.1109/JSTSP.2007.910281).
12. 12)
  - 2. Wagner, A., Wright, J., Ganesh, A., Zhou, Z., Mobahi, H., Ma, Y.: ‘Towards a practical face recognition system: Robust alignment and illumination via sparse representation’, IEEE Trans. Pattern Anal. Mach. Intell., 2012, 34, (2), pp. 372–386 (doi: 10.1109/TPAMI.2011.112).
13. 13)
  - 107. Georghiades, A., Belhumeur, P., 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 (doi: 10.1109/34.927464).
14. 14)
  - 22. Wang, H., Zheng, W.: ‘Robust sparsity-preserved learning with application to image visualization’, Knowl. Inf. Syst., 2014, 39, (2), pp. 287–304 (doi: 10.1007/s10115-012-0605-7).
15. 15)
  - 8. Zhang, S., Yao, H., Sun, X., Lu, X.: ‘Sparse coding based visual tracking: Review and experimental comparison’, Pattern Recognit., 2013, 46, (7), pp. 1772–1788 (doi: 10.1016/j.patcog.2012.10.006).
16. 16)
  - 17. Zhao, Z.-Q., Glotin, H., Xie, Z., Gao, J., Wu, X.: ‘Cooperative sparse representation in two opposite directions for semi-supervised image annotation’, IEEE Trans. Image Process., 2012, 21, (9), pp. 4218–4231 (doi: 10.1109/TIP.2012.2197631).
17. 17)
  - 37. Deng, W., Hu, J., Guo, J.: ‘Extended SRC: undersampled face recognition via intraclass variant dictionary’, IEEE Trans. Pattern Anal. Mach. Intell., 2012, 34, (9), pp. 1864–1870 (doi: 10.1109/TPAMI.2012.30).
18. 18)
  - 13. Donoho, D.L., Tsaig, Y.: ‘Fast solution of ℓ1-norm minimization problems when the solution may be sparse’, IEEE Trans. Inf. Theory, 2008, 54, (11), pp. 4789–4812 (doi: 10.1109/TIT.2008.929958).
19. 19)
  - 12. Yang, J., Zhang, Y.: ‘Alternating direction algorithms for ℓ1-problems in compressive sensing’, SIAM J. Sci. Comput., 2011, 33, (1), pp. 250–278 (doi: 10.1137/090777761).
20. 20)
  - 5. Zhao, Z.-Q., Huang, D.S., Sun, B.-Y.: ‘Human face recognition based on multiple features using neural networks committee’, Pattern Recogn. Lett., 2004, 25, (12), pp. 1351–1358 (doi: 10.1016/j.patrec.2004.05.008).
21. 21)
  - 26. Yang, A.Y., Sastry, S.S., Ganesh, A., Ma, Y.: ‘Fast ℓ1at-minimization algorithms and an application in robust face recognition: A review’. ICIP, 2010, pp. 1849–1852.
22. 22)
  - 16. Ho, J., Yang, M., Lim, J., Lee, K., Kriegman, D.: ‘Clustering appearances of objects under varying illumination conditions’. CVPR, 2003, pp. 11–18.
23. 23)
  - 10. Monteiro, R.D., Adler, I.: ‘Interior path following primal-dual algorithms. part i: Linear programming’, Math. Program, 1989, 44, (1), pp. 27–41 (doi: 10.1007/BF01587075).
24. 24)
  - 9. Efron, B., Hastie, T., Johnstone, I., Tibshirani, R.: ‘Least angle regression’, Ann. Statist. (With Discussion), 2004, 32, pp. 407–499 (doi: 10.1214/009053604000000067).
25. 25)
  - 13. He, X., Yan, S., Hu, Y., et al: ‘Face recognition using Laplacianface’, IEEE Trans. Pattern Anal. Mach. Intell., 2005, 27, (3), pp. 328–340 (doi: 10.1109/TPAMI.2005.55).
26. 26)
  - 108. Lee, K.-C., Ho, J., Kriegman, D.: ‘Acquiring linear subspaces for face recognition under variable lighting’, IEEE Trans. Pattern Anal. Mach. Intell., 2005, 27, (5), pp. 684–698 (doi: 10.1109/TPAMI.2005.92).
27. 27)
  - 9. Shi, Q., Eriksson, A., van den Hengel, A., Shen, C.: ‘Is face recognition really a compressive sensing problem?’. CVPR, 2011, pp. 553–560.
28. 28)
  - 23. Kojima, M., Megiddo, N., Mizuno, S.: ‘Theoretical convergence of large-step primal-dual interior point algorithms for linear programming’, Math. Program, 1993, 59, pp. 1–21 (doi: 10.1007/BF01581234).
29. 29)
  - 9. Zhang, S., Yao, H., Zhou, H., Sun, X., Liu, S.: ‘Robust visual tracking based on online learning sparse representation’, Neurocomputing, 2013, 100, pp. 31–40 (doi: 10.1016/j.neucom.2011.11.031).
30. 30)
  - 32. Martínez, A., Benavente, R.: ‘The ar face database’, June 1998, http://www2.ece.ohio-state.edu/aleix/ARdatabase.html.
31. 31)
  - 14. Lu, C.-Y., Min, H., Zhao, Z.-Q., Zhu, L., Huang, D.-S., Yan, S.: ‘Robust and efficient subspace segmentation via least squares regression’. ECCV, 2012, pp. 347–360.
32. 32)
  - 28. http://www.eecs.berkeley.edu/~yang/software/l1benchmark/l1benchmark.zip, accessed: 2013-10-09.

Errata

An Erratum has been published for this content:

Erratum: “Expanding dictionary for robust face recognition: pixel is not necessary while sparsity is”

Expanding dictionary for robust face recognition: pixel is not necessary while sparsity is

References

Related content