A Hybrid Face Recognition Approach Using GPUMLib

  • Noel Lopes
  • Bernardete Ribeiro
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6419)

Abstract

We present a hybrid face recognition approach which relies on a Graphics Processing Unit (GPU) Machine Learning (ML) Library (GPUMLib). The library includes a high-performance implementation of the Non-Negative Matrix Factorization (NMF) and the Multiple Back-Propagation (MBP) algorithms. Both algorithms are combined in order to obtain a reliable face recognition classifier. The proposed approach first applies an histogram equalization to the original face images in order to reduce the influence from the surrounding illumination. The NMF algorithm is then applied to reduce the data dimensionality, while preserving the information of the most relevant features. The obtained decomposition is further used to rebuild accurate approximations of the original data (by using additive combinations of the parts-based matrix). Finally, the MBP algorithm is used to build a neural classifier with great potential to construct a generalized solution. Our approach is tested in the Yale face database, yielding an accuracy of 93.33% thus demonstrating its potential. Moreover, the speedups obtained with the GPU greatly enhance real-time implementation face recognition systems.

Keywords

GPU computing Non-Negative Matrix Factorization Multiple Back-Propagation Hybrid systems Face Recognition 

References

  1. 1.
    Gillis, N., Glineur, F.: Using underapproximations for sparse nonnegative matrix factorization. Pattern Recogntion 43(4), 1676–1687 (2010)CrossRefMATHGoogle Scholar
  2. 2.
    Lee, D.D., Seung, H.S.: Learning the parts of objects by non-negative matrix factorization. Nature 401, 788–791 (1999)CrossRefMATHGoogle Scholar
  3. 3.
    Lopes, N., Ribeiro, B.: An efficient gradient-based learning algorithm applied to neural networks with selective actuation neurons. Neural, Parallel & Scientific Computations 11(3), 253–272 (2003)MathSciNetMATHGoogle Scholar
  4. 4.
    Lopes, N., Ribeiro, B.: GPU implementation of the multiple back-propagation algorithm. In: Corchado, E., Yin, H. (eds.) IDEAL 2009. LNCS, vol. 5788, pp. 449–456. Springer, Heidelberg (2009)CrossRefGoogle Scholar
  5. 5.
    Lopes, N., Ribeiro, B.: GPUMLib: a new library to combine machine learning algorithms with graphics processing units. In: 10th International Conference on Hybrid Intelligent Systems, Atlanta, USA (2010)Google Scholar
  6. 6.
    Lopes, N., Ribeiro, B.: Non-negative matrix factorization implementation using graphic processing units. In: Fyfe, C., Tino, P., Charles, D., Garcia-Osomo, C., Yin, H. (eds.) IDEAL 2010. LNCS, vol. 6283, pp. 275–283. Springer, Heidelberg (2010)Google Scholar
  7. 7.
    Tang, H., Tan, K.C., Yi, Z.: Neural Networks: Computational Models and Applications. Studies in Computational Intelligence. Springer, New York (2007)CrossRefMATHGoogle Scholar
  8. 8.
    Wang, J., Zhang, B., Wang, S., Qi, M., Kong, J.: An adaptively weighted sub-pattern locality preserving projection for face recognition. Journal of Network and Computer Applications 33(3), 323–332 (2010)CrossRefGoogle Scholar
  9. 9.
    Xu, B., Lu, J., Huang, G.: A constrained non-negative matrix factorization in information retrieval. In: IEEE International Conference on Information Reuse and Integration, IRI 2003, pp. 273–277 (2003)Google Scholar
  10. 10.
    Zhao, W., Chellappa, R., Phillips, P.J., Rosenfeld, A.: Face recognition: A literature survey. ACM Computing Surveys 35(4), 399–458 (2003)CrossRefGoogle Scholar
  11. 11.
    Zilu, Y., Guoyi, Z.: Facial expression recognition based on NMF and SVM. International Forum on Information Technology and Applications 3, 612–615 (2009)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  • Noel Lopes
    • 1
    • 2
  • Bernardete Ribeiro
    • 1
  1. 1.CISUC - Center for Informatics and Systems of University of CoimbraPortugal
  2. 2.UDI/IPG - Research UnitPolytechnic Institute of GuardaPortugal

Personalised recommendations