Joint Diagonalization of Kernels for Information Fusion

  • Alberto Muñoz
  • Javier González
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4756)


Information Fusion is becoming increasingly relevant in fields such as Image Processing or Information Retrieval. In this work we propose a new technique for information fusion when the sources of information are given by a set of kernel matrices. The algorithm is based on the joint diagonalization of matrices and it produces a new data representation in an Euclidean space. In addition, the proposed method is able to eliminate redundant information among the input kernels and it is robust against the presence of noisy variables and irrelevant kernels.

The performance of the algorithm is illustrated on data reconstruction and classifications problems.


Information Fusion Approximate Joint Diagonalization Kernel Methods Support Vector Machines 


  1. 1.
    Bach, F.R., Jordan, M.I.: Kernel Principal Components Analysis. Journal of Machine Learning Research 3, 1–48 (2002)CrossRefMathSciNetGoogle Scholar
  2. 2.
    Cardoso, J.F., Souloumiac: A Jacobi Angles for Simultaneous Diagonalization. SIAM J. Mat. Anals. Applied 17(1), 161–164 (1996)zbMATHCrossRefMathSciNetGoogle Scholar
  3. 3.
    Choi, M., Young, R., Nam, M.-R., Kim, H.O.: Fusion of Multispectral and Panchromatic Satellite Images Using the Curvelet Transform. Geoscience and Remote Sensing Letters 2(2) (2005)Google Scholar
  4. 4.
    Martin de Diego, I., Moguerza, J.M., Muñoz, A.: Combining Kernel Information for Support Vector Classification. In: Roli, F., Kittler, J., Windeatt, T. (eds.) MCS 2004. LNCS, vol. 3077, pp. 102–111. Springer, Heidelberg (2004)Google Scholar
  5. 5.
    Epifanio, I., Gutierrez, J., Malo, J.: Linear transform for simultaneous diagonalization of covariance and perceptual metric matrix in image coding. Pattern Recognition 36, 799–1811 (2003)CrossRefGoogle Scholar
  6. 6.
    Gretton, A., Herbrich, R., Smola, A., Bousquet, O., Schölkop, B.: Kernel methods for measuring independence. J. Machine Learning Research 6, 2075–2129 (2005)Google Scholar
  7. 7.
    Hua, Y.: On SVD estimating Generalized Eigenvalues of Singular Matrix Pencils in Noise. IEEE Transactions on Signal Processing 39(4), 892–900 (1991)CrossRefMathSciNetGoogle Scholar
  8. 8.
    Joachims, T., Cristianini, N., Shawe-Taylor, J.: Composite Kernels for Hypertext Categorisation. In: Proceedings of the International Conference on Machine Learning, pp. 250–257 (2002)Google Scholar
  9. 9.
    Kittler, J., Hatef, M., Duin, R.P.W., Matas, J.: On combining classifiers. IEEE Transactions on Pattern Analysis and Machine intelligence 20(3), 226–239 (1998)CrossRefGoogle Scholar
  10. 10.
    Mika, S., Ratsch, W.G., Scholkopf, B., Muller, K.-R.: Fisher Discriminant Analysis with Kernels. In: Proceedings of IEEE Neural Networks for Signal Processing Workshop, IEEE Computer Society Press, Los Alamitos (1999)Google Scholar
  11. 11.
    Moguerza, J., Muñoz, A.: Support Vector Machines with Applications. Statistical Science 21(3), 322–336 (2006)CrossRefGoogle Scholar
  12. 12.
    Beresford, P.N.: The Symmetric Eigenvalue Problem. Classics in Applied Mathematics. SIAM (1997)Google Scholar
  13. 13.
    Schölkopf, B., Smola, A.J., Müller, K.R.: Nonlinear Component Analysis as a Kernel Eigenvalue Problem. Neural Computation 10, 1299–1319 (1998)CrossRefGoogle Scholar
  14. 14.
    Wax, M., Sheinvald, J.: A least-squares approach to joint diagonalization. IEEE Signal Processing Lett. 4, 52–53 (1997)CrossRefGoogle Scholar
  15. 15.
    Bunse-Gerstner, A., Byers, R., Mehrmann, V.: Numerical methods for simultaneous diagonalization. SIAM Journal on Matrix Analysis and Applications 14(4), 927–949 (1993)zbMATHCrossRefMathSciNetGoogle Scholar
  16. 16.
    Yeredor, A.: Non-Orthogonal Joint Diagonalization in the Least-Squares Sense With Application in Blind Source Separation. IEEE Transactions on Signal Processing. 50 (7), 1545–1553 (2002)CrossRefMathSciNetGoogle Scholar
  17. 17.
    Piella, G., Heijmans, H.: Multiresolution Image Fusion Guided by a Multimodal Segmentation. In: Proceedings of ACIVS 2002, Ghent, Belgium (September 9-11, 2002)Google Scholar
  18. 18.
    Newman, D.J., Hettich, S., Blake, C.L. and Merz, C.J.UCI Repository of machine learning databases. Irvine, CA: University of California, Department of Information and Computer Science 1998,

Copyright information

© Springer-Verlag Berlin Heidelberg 2007

Authors and Affiliations

  • Alberto Muñoz
    • 1
  • Javier González
    • 1
  1. 1.Universidad Carlos III de Madrid, c/ Madrid 126, 28903 GetafeSpain

Personalised recommendations