Optimal Learning High-Order Markov Random Fields Priors of Colour Image

  • Ke Zhang
  • Huidong Jin
  • Zhouyu Fu
  • Nianjun Liu
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4843)


In this paper, we present an optimised learning algorithm for learning the parametric prior models for high-order Markov random fields (MRF) of colour images. Compared to the priors used by conventional low-order MRFs, the learned priors have richer expressive power and can capture the statistics of natural scenes. Our proposed optimal learning algorithm is achieved by simplifying the estimation of partition function without compromising the accuracy of the learned model. The parameters in MRF colour image priors are learned alternatively and iteratively in an EM-like fashion by maximising their likelihood. We demonstrate the capability of the proposed learning algorithm of high-order MRF colour image priors with the application of colour image denoising. Experimental results show the superior performance of our algorithm compared to the state–of–the–art of colour image priors in [1], although we use a much smaller training image set.


Markov random fields image prior colour image denoising 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    McAuley, J., Caetano, T., Smola, A., Franz, M.: Learning high-order MRF priors of color images. In: ICML. LNCS, vol. 4503, pp. 617–624. Springer, Heidelberg (2006)CrossRefGoogle Scholar
  2. 2.
    Freeman, W., Pasztor, E., Carmichael, O.: Learning low-level vision. International Journal of Computer Vision 40, 25–47 (2000)zbMATHCrossRefGoogle Scholar
  3. 3.
    Roth, S., Black, M.: Fields of experts: A framework for learning image priors. In: ICCV, pp. 860–867 (2005)Google Scholar
  4. 4.
    Olshausen, B., Field, D.: Sparse coding with an overcomlete basis set: a strategy employed by v1? Vision Research 37, 3311–3325 (1997)CrossRefGoogle Scholar
  5. 5.
    Welling, M., Hinton, G., Osindero, S.: Learning sparse topographic representations with products of Student-t distributions. In: NIPS, vol. 15, pp. 1359–1366 (2003)Google Scholar
  6. 6.
    Geman, S., Geman, D.: Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images. PAMI 6(6), 721–741 (1984)zbMATHGoogle Scholar
  7. 7.
    Hinton, G.: Products of experts. In: 9th ICANN, pp. 1–6 (1999)Google Scholar
  8. 8.
    Zhu, S., Wu, Y., Mumford, D.: Filters, random fields and maximum entropy (frame): Towards a unified theory of texture modeling. International Journal of Computer Vision 27, 107–126 (1998)CrossRefGoogle Scholar
  9. 9.
    Abramowitz, M., Stegun, I.E.: The process of the Arithmetic–Geometric mean. Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th printing, 571 (1972)Google Scholar
  10. 10.
    Krantz, S.: In: Handbook of Complex Variables. Boston, MA:Brikhauser, p.118 (1999)Google Scholar
  11. 11.
    Moré, J., Thuente, D.: Line search algorithms with guaranteed sufficient decrease. ACM Trans. Math. Software 20, 286–307 (1994)zbMATHCrossRefMathSciNetGoogle Scholar
  12. 12.
    Martin, D., Fowlkes, C., Tal, D., Malik, J.: A database of human segmented natural images and ites application to evaluating segmentation algorithms and measuring ecological statistics. In: 8th ICCV, pp. 416–423 (2001)Google Scholar
  13. 13.
  14. 14.
    Portilla, J., Strela, V., Wainwright, M., Simoncelli, E.: Image denoising using scale mixtures of Gaussians in the wavelet domain. IEEE Trans. Image Processing. 12(11), 1338–1351 (2003)CrossRefMathSciNetGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2007

Authors and Affiliations

  • Ke Zhang
    • 1
    • 2
  • Huidong Jin
    • 1
    • 2
  • Zhouyu Fu
    • 1
    • 2
  • Nianjun Liu
    • 1
    • 2
  1. 1.Research School of Information Sciences and Engineering (RSISE), Australian National University 
  2. 2.National ICT Australia (NICTA), Canberra Lab, ACTAustralia

Personalised recommendations