Markov–Gibbs Texture Modelling with Learnt Freeform Filters

  • Ralph VersteegenEmail author
  • Georgy Gimel’farb
  • Patricia Riddle
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10029)


Energy-based Markov–Gibbs random field (MGRF) image models describe images by statistics of localised features; hence selecting statistics is crucial. This paper presently a procedure for searching much broader than typical families of linear-filter-based statistics, by alternately optimising in continuous parameter space and discrete graphical structure space. This unifies and extends the divergent models deriving from the well-known Fields of Experts (FoE), which learn parametrised features built on small linear filters, and the constrasting FRAME (exponential family) approach which iteratively selects large filters from a fixed set. While FoE is limited by computational cost to small filters, we use large sparse (non-contiguous) filters with arbitrary shapes which can capture long-range interactions directly. A filter pre-training step also improves speed and results. Synthesis of a variety of textures shows promising abilities of the proposed models to capture both fine details and larger-scale structure with a low number of small and efficient filters.


Local Binary Pattern Gibbs Sampling Nest Iteration Texture Synthesis Texture Model 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


  1. 1.
    Barndorff-Nielsen, O.: Information and Exponential Families in Statistical Theory. Wiley, Hoboken (1978)zbMATHGoogle Scholar
  2. 2.
    Della Pietra, S., Della Pietra, V., Lafferty, J.: Inducing features of random fields. IEEE Trans. Pattern Anal. Mach. Intell. 19(4), 380–393 (1997)CrossRefGoogle Scholar
  3. 3.
    Gao, Q., Roth, S.: How well do filter-based MRFs model natural images? In: Pinz, A., Pock, T., Bischof, H., Leberl, F. (eds.) DAGM/OAGM 2012. LNCS, vol. 7476, pp. 62–72. Springer, Heidelberg (2012). doi: 10.1007/978-3-642-32717-9_7 CrossRefGoogle Scholar
  4. 4.
    Gao, Q., Roth, S.: Texture synthesis: from convolutional RBMs to efficient deterministic algorithms. In: Fränti, P., Brown, G., Loog, M., Escolano, F., Pelillo, M. (eds.) S+SSPR 2014. LNCS, vol. 8621, pp. 434–443. Springer, Heidelberg (2014). doi: 10.1007/978-3-662-44415-3_44 Google Scholar
  5. 5.
    Gatys, L.A., Ecker, A.S., Bethge, M.: Texture synthesis using convolutional neural networks. In: Advances in Neural Information Processing Systems, pp. 262–270 (2015)Google Scholar
  6. 6.
    Gimel’farb, G.: Image Textures and Gibbs Random Fields. Kluwer Academic Publishers, Dordrecht (1999)CrossRefzbMATHGoogle Scholar
  7. 7.
    Hao, T., Raiko, T., Ilin, A., Karhunen, J.: Gated Boltzmann machine in texture modeling. In: Villa, A.E.P., Duch, W., Érdi, P., Masulli, F., Palm, G. (eds.) ICANN 2012. LNCS, vol. 7553, pp. 124–131. Springer, Heidelberg (2012). doi: 10.1007/978-3-642-33266-1_16 CrossRefGoogle Scholar
  8. 8.
    Heess, N., Williams, C.K.I., Hinton, G.E.: Learning generative texture models with extended fields-of-experts. In: Proceedings of British Machine Vision Conference (BMVC 2009), pp. 1–11 (2009)Google Scholar
  9. 9.
    Hinton, G.E.: Training products of experts by minimizing contrastive divergence. Neural Comput. 14(8), 1771–1800 (2002)MathSciNetCrossRefzbMATHGoogle Scholar
  10. 10.
    Kivinen, J.J., Williams, C.: Multiple texture Boltzmann machines. In: Proceedings of 15th International Conference on Artificial Intelligence and Statistics, vol. 2, pp. 638–646 (2012)Google Scholar
  11. 11.
    Köster, U., Lindgren, J.T., Hyvärinen, A.: Estimating Markov random field potentials for natural images. In: Adali, T., Jutten, C., Romano, J.M.T., Barros, A.K. (eds.) ICA 2009. LNCS, vol. 5441, pp. 515–522. Springer, Heidelberg (2009). doi: 10.1007/978-3-642-00599-2_65 CrossRefGoogle Scholar
  12. 12.
    Lee, S., Ganapathi, V., Koller, D.: Efficient structure learning of Markov networks using L1-regularization. In: Advances in Neural Information Processing Systems 19 (NIPS 2006), pp. 817–824 (2007)Google Scholar
  13. 13.
    Luo, H., Carrier, P.L., Courville, A., Bengio, Y.: Texture modeling with convolutional spike-and-slab RBMs and deep extensions. J. Mach. Learn. Res. 31, 415–423 (2013)Google Scholar
  14. 14.
    Peyré, G.: Sparse modeling of textures. J. Math. Imaging Vis. 34(1), 17–31 (2009)MathSciNetCrossRefGoogle Scholar
  15. 15.
    Portilla, J., Simoncelli, E.P.: A parametric texture model based on joint statistics of complex wavelet coefficients. Int. J. Comput. Vis. 40(1), 49–70 (2000)CrossRefzbMATHGoogle Scholar
  16. 16.
    Ranzato, M., Mnih, V., Hinton, G.E.: Generating more realistic images using gated MRFs. In: Advances in Neural Information Processing Systems 23 (NIPS 2010), pp. 2002–2010. Curran Associates (2010)Google Scholar
  17. 17.
    Roth, S., Black, M.J.: Fields of experts. Int. J. Comput. Vis. 82(2), 205–229 (2009)CrossRefGoogle Scholar
  18. 18.
    Schmidt, U., Gao, Q., Roth, S.: A generative perspective on MRFs in low-level vision. In: IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2010), pp. 1751–1758 (2010)Google Scholar
  19. 19.
    Tieleman, T.: Training restricted Boltzmann machines using approximations to the likelihood gradient. In: Proceedings of 25th International Conference on Machine Learning (ICML 2008), pp. 1064–1071. ACM (2008)Google Scholar
  20. 20.
    Versteegen, R., Gimel’farb, G., Riddle, P.: Texture modelling with non-contiguous filters. In: Proceedings of 30th International Conference on Image and Vision Computing New Zealand (IVCNZ 2015) (2015)Google Scholar
  21. 21.
    Versteegen, R., Gimel’farb, G., Riddle, P.: Texture modelling with nested high-order Markov-Gibbs random fields. Comput. Vis. Image Underst. 143, 120–134 (2016)CrossRefGoogle Scholar
  22. 22.
    Zhu, S.C., Wu, Y., Mumford, D.: Minimax entropy principle and its application to texture modeling. Neural Comput. 9(8), 1627–1660 (1997)CrossRefGoogle Scholar
  23. 23.
    Zhu, S.C., Wu, Y., Mumford, D.: Filters, random fields and maximum entropy (FRAME): towards a unified theory for texture modeling. Int. J. of Comput. Vis. 27(2), 107–126 (1998)CrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG 2016

Authors and Affiliations

  • Ralph Versteegen
    • 1
    Email author
  • Georgy Gimel’farb
    • 1
  • Patricia Riddle
    • 1
  1. 1.Department of Computer ScienceThe University of AucklandAucklandNew Zealand

Personalised recommendations