Unsupervised Image Segmentation Using Markov Random Fields

  • Abdulkadir Şengür
  • İbrahim Türkoğlu
  • M. Cevdet İnce
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3949)


In this study, we carried out an unsupervised gray level image segmentation based on Markov Random Fields (MRF) model. First, we use the Expectation Maximization (EM) algorithm to estimate the distribution of the input image and the number of the components is automatically determined by the Minimum Message Length (MML) algorithm. Then the segmentation is done by the Iterated Conditional Modes (ICM) algorithm. For testing the segmentation performance, we use both artificial images and real images. The experimental results are satisfactory.


Image Segmentation Real Image Markov Random Fields Expectation Maximization Algorithm True Image 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Gonzalez, R.C., Woods, R.E.: Digital image processing. Prentice Hall, Englewood Cliffs (2002)Google Scholar
  2. 2.
    Sahoo, P.K., Soltani, S., Wong, A., Chen, Y.: A survey of thresholding techniques. Computer Vision Graphics Image Processing 41, 233–260 (1988)CrossRefGoogle Scholar
  3. 3.
    Şengür, A., Türkoğlu, İ., İnce, M.C.: Performance Comparison of Thresholding Algorithms on Uneven Illuminated Image. Asian Journal of Information technology 3(10), 956–959 (2004)Google Scholar
  4. 4.
    Tsao, E.C.K., Bezdek, J.C., Pal, N.R.: Fuzzy Kohonen clustering networks. Patt. Recog. 27, 757–764 (1994)CrossRefGoogle Scholar
  5. 5.
    Adams, R., Bischof, L.: Seeded region growing. IEEE Trans. on PAMI, 641–647 (1994)Google Scholar
  6. 6.
    Kurugollu, F., Sankur, B., Harmanci, A.E.: Color image segmentation using histogram multithresholding and fusion. Image and Vision Computing 19, 915–928 (2001)CrossRefGoogle Scholar
  7. 7.
    Geman, S., Geman, D.: Stochastic relaxation, Gibbs distributions, and Bayesian restoration of images. IEEE Trans. PAMI 6, 721–741 (1984)CrossRefMATHGoogle Scholar
  8. 8.
    Deng, H., Clausi, D.A.: Unsupervised image segmentation using a simple MRF model with a new implementation schema. Patt. Recog. 37, 2322–2335 (2004)CrossRefGoogle Scholar
  9. 9.
    Besag, J.: On the statistical analysis of dirty pictures. J. Roy. Statist Soc. Ser. 48, 259–302 (1986)MathSciNetMATHGoogle Scholar
  10. 10.
    Kato, Z., Zerubia, J., Berthod, M.: Unsupervised parallel image classification using Markovian models. Patt. Recog. 32, 591–604 (1999)CrossRefGoogle Scholar
  11. 11.
    Kato, Z., Pong, T.C., Lee, J.: Color image segmentation and parameter estimation in a markovian framework. Patt. Recog. Lett. 22, 309–321 (2001)CrossRefMATHGoogle Scholar
  12. 12.
    Yang, X., Liu, J.: Unsupervised texture segmentation with one-step mean shift and boundary Markov random fields. Pattern Recognition Letters 22, 1073–1081 (2001)CrossRefMATHGoogle Scholar
  13. 13.
    Dempster, A., Laird, N., Rubin, D.: Maximum likelihood estimation from incomplete data via the EM algorithm. J. Royal Statistical Soc. B. 39, 1–38 (1977)MathSciNetMATHGoogle Scholar
  14. 14.
    Wallace, C., Dowe, D.: Minimum Message Length and Kolmogorov complexity. The computer J. 42, 270–283 (1999)CrossRefMATHGoogle Scholar
  15. 15.
    Dubes, R.C., Jain, A.K., Nadabar, S.G., Chen, C.C.: MRF model based algorithms for image segmentation. IEEE Computer Society Press, Los Alamitos (1990)CrossRefGoogle Scholar
  16. 16.
    McLanchlan, G., Krishnan, T.: The EM algorithm and extensions. John Wiley and Sons, New York (1997)Google Scholar
  17. 17.

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • Abdulkadir Şengür
    • 1
  • İbrahim Türkoğlu
    • 1
  • M. Cevdet İnce
    • 2
  1. 1.Department of Electronic and Computer ScienceFırat UniversityElazı
  2. 2.Department of Electric-Electronic EngineeringFırat UniversityElazı

Personalised recommendations