Accelerated Nonlinear Gaussianization for Feature Extraction

Alexandru Paul Condurache, Alfred Mertins

2013

Abstract

In a multi-class classification setup, the Gaussianization represents a nonlinear feature extraction transform with the purpose of achieving Gaussian class-conditional densities in the transformed space. The computational complexity of such a transformation increases with the dimension of the processed feature space in such a way that only relatively small dimensions can be processed. In this contribution we describe how to reduce the computational burden with the help of an adaptive grid. Thus, the Gaussianization transform is able to also handle feature spaces of higher dimensionality, improving upon its practical usability. On both artificially generated and real-application data, we demonstrate a decrease in computation complexity in comparison to the standard Gaussianization, while maintaining the effectiveness.

References

  1. Bezdek, J., Keller, J., Krishnapuram, R., Kuncheva, L., and Pal, N. (1999). Will the real iris data please stand up? IEEE Trans. on Fuzzy Systems, 7(3):368-369.
  2. Bishop, C. M. (2009). Pattern recognition and machine learning. Information Science and Statistics. Springer.
  3. Chen, S. and Gopinath, R. (2000). Gaussianization. In Proc. of NIPS, Denver, USA.
  4. Condurache, A. P. and Mertins, A. (2011). Elastictransform based multiclass gaussianization. IEEE Signal Processing Letters, 18(8):482-485.
  5. Dias, T. M., Attux, R., Romano, J. M., and Suyama, R. (2009). Blind source separation of post-nonlinear mixtures using evolutionary computation and gaussianization. In Proc. of ICA, pages 235-242. Springer.
  6. Gonzales, R. C. and Woods, R. E. (2008). Digital image processing (Third edition). Pearson Education.
  7. Gopinath, R. A. (1998). Maximum likelihood modeling with gaussian distributions for classification. In Proc. of ICASSP, pages 661-664, Seattle, U.S.A.
  8. Hackbusch, W. (1993). Iterative solution of large sparse systems of equations. Springer-Verlag (New York).
  9. Hogg, R., Craig, A., and McKean, J. (2004). Introduction to mathematical statistics. Prentice Hall, 6 edition.
  10. Mezghani-Marrakchi, I., Mah, G., Jadane-Sadane, M., Djaziri-Larbi, S., and Turki-Hadj-Allouane, M. (2007). ”gaussianization” method for identification of memoryless nonlinear audio systems. In Proc. of EUSIPCO, pages 2316 - 2320.
  11. Modersitzki, J. (2004). Numerical Methods for Image Registration. Oxford University Press.
  12. Rachev, S. T. (1985). The Monge-Kantorovich mass transference problem and its stochastic applications. SIAM Theory of Probability and its Applications, 29(4):647 - 676.
  13. Saon, G., Dharanipragada, S., and Povey, D. (2004). Feature space gaussianization. In Proc. of ICASSP, pages I - 329-332.
  14. Trottenberg, U., Oosterlee, C. W., and Schller, A. (2001). Multigrid. Academic Press.
  15. Wesseling, P. (1992). An introduction to multigrid methods. John Wiley & Sons.
Download


Paper Citation


in Harvard Style

Paul Condurache A. and Mertins A. (2013). Accelerated Nonlinear Gaussianization for Feature Extraction . In Proceedings of the 2nd International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM, ISBN 978-989-8565-41-9, pages 121-126. DOI: 10.5220/0004204701210126


in Bibtex Style

@conference{icpram13,
author={Alexandru Paul Condurache and Alfred Mertins},
title={Accelerated Nonlinear Gaussianization for Feature Extraction},
booktitle={Proceedings of the 2nd International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,},
year={2013},
pages={121-126},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0004204701210126},
isbn={978-989-8565-41-9},
}


in EndNote Style

TY - CONF
JO - Proceedings of the 2nd International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,
TI - Accelerated Nonlinear Gaussianization for Feature Extraction
SN - 978-989-8565-41-9
AU - Paul Condurache A.
AU - Mertins A.
PY - 2013
SP - 121
EP - 126
DO - 10.5220/0004204701210126