This Article 
 Bibliographic References 
 Add to: 
An Empirical Study of the Simulation of Various Models used for Images
May 1994 (vol. 16 no. 5)
pp. 507-513

Markov random fields are typically used as priors in Bayesian image restoration methods to represent spatial information in the image. Commonly used Markov random fields are not in fact capable of representing the moderate-to-large scale clustering present in naturally occurring images and can also be time consuming to simulate, requiring iterative algorithms which can take hundreds of thousands of sweeps of the image to converge. Markov mesh models, a causal subclass of Markov random fields, are, however, readily simulated. We describe an empirical study of simulated realizations from various models used in the literature, and we introduce some new mesh-type models. We conclude, however, that while large-scale clustering may be represented by such models, strong directional effects are also present for all but very limited parameterizations. It is emphasized that the results do not detract from the use of Markov random fields as representers of local spatial properties, which is their main purpose in the implementation of Bayesian statistical approaches to image analysis. Brief allusion is made to the issue of parameter estimation.

[1] K. Abend, T. J. Harley, and L. N. Kanal, "Classification of binary random patterns,"IEEE Trans. Inform. Theory, vol. IT-4, pp. 538-544, 1965.
[2] N. Ahuja and A. Rosenfeld, "Mosaic models for textures,"IEEE Trans. Pattern Anal. Machine Intell., vol. PAMI-3, pp. 1-11, 1981.
[3] N. Ahuja and B. J. Schachter, "Image models,"Computing Surveys. vol. 13, pp. 374-397, 1981.
[4] M. S. Bartlett,The Statistical Analysis of Spatial Pattern. London: Chapman and Hall, 1975.
[5] R. J. Baxter, "Partition function of the eight-vertex lattice model,"Ann. Phys., vol. 70, 193-228, 1970.
[6] J. Besag, "Spatial interaction and the statistical analysis of lattice systems (with Discussion),"J. Roy. Statist. Soc. B, vol. 36, pp. 192-236, 1974.
[7] J. Besag, "Statistical analysis of non-lattice data,"The Statistician, vol. 24, pp. 179-195, 1975.
[8] J. Besag, "On the statistical analysis of dirty pictures (with discussion),"J. Roy. Statist. Soc. B, vol. 48, pp. 259-302, 1986.
[9] J. Besag and P. J. Green, "Spatial statistics and Bayesian computation,"J. Roy. Statist. Soc. B, vol. 55, pp. 25-37, 1993.
[10] C. C. Chen, "Markov random field models in image analysis," Ph.D. thesis, Michigan State Univ., East Lansing, MI, 1980.
[11] G. R. Cross and A. K. Jain, "Markov random field texture models,"IEEE Trans. Pattern. Anal. Machine Intell., vol. PAMI-5, pp. 25-39, 1983.
[12] H. Derin and H. Elliott, "Modeling and segmentation of noisy and textured images using Gibbs random fields,"IEEE Trans. Pattern Anal. Machine Intell., vol. PAMI-9, pp. 39-55, Jan. 1987.
[13] P. A. Devijver, "Image segmentation using causal Markov random field models," inProc. 4th Int. Conf. Pattern Recognit., 1988, pp. 131-143.
[14] P. A. Devijver, "Real-time modelling of image sequences based on hidden Markov mesh random field models," Manuscript M-307, Phillips Res. Lab., Brussels, Belgium, 1989.
[15] R. C. Dubes and A. K. Jain, "Random field models in image analysis,"J. Appl. Statist., vol. 16, pp. 131-164, 1989.
[16] P. A. Flinn, "Monte-Carlo calculation of phase separation in a two-dimensional Ising system,"J. Statist. Phys., vol. 10, pp. 89-97, 1974.
[17] S. Geman and D. Geman, "Stochastic relaxation, Gibbs distributions and the Bayesian restoration of images,"IEEE Trans. Pattern. Anal. Machine Intell., vol. PAMI-6, pp. 721-741, 1984.
[18] A. J. Gray, "Simulating posterior Gibbs distributions: a comparison of the Swendsen-Wang and Gibbs sampler methods,"Statist. Computing, to appear, 1994.
[19] P. J. Green and X.-L. Han, "Metropolis methods, gaussian proposals and antithetic variables," inLecture Notes in Statistics, vol. 74, P. Baroneet al., Eds. Berlin: Springer, 1992, pp. 142-164.
[20] D. M. Greig, B. T. Porteous, and A. H. Seheult, "Exact maximum aposterioriestimation for binary images,"J. Roy. Statist. Soc. B, vol. 51, pp. 271-279, 1989.
[21] E. Ising, "Beitrag zur theorie des ferromagnetismus,"Zeitschrift Physik, vol. 31, p. 253, 1925.
[22] L. N. Kanal, "Markov mesh models,"Computer Graphics and Image Processing, vol. 12, pp. 371-375, 1980.
[23] N. Metropolis, A. W. Rosenbluth, M. N. Rosenbluth, A. H. Teller, and E. Teller, "Equations of state calculations by fast computing machines,"J. Chem. Phys., vol. 21, pp. 1087-1092, 1953.
[24] G. F. Newell and E. W. Montroll, "On the theory of the Ising model of ferromagnetism,"Revs. Modern Phys., vol. 25, pp. 353-389, 1953.
[25] L. Onsager, "Crystal Statistics I: A two-dimensional model with an order-disorder transition,"Phys. Rev., vol. 65, pp. 117-149, 1944.
[26] D. K. Pickard, "Statistical inference on lattices," Unpublished Ph.D. thesis, Australian Nat. Univ., 1977.
[27] D. K. Pickard, "A curious binary lattice process,"J. Appl. Probab., vol. 14, pp. 717-731, 1977.
[28] D. K. Pickard, "Unilateral Ising models,"Suppl. Adv. Appl. Probab., vol. 10, pp. 58-64, 1978.
[29] D. K. Pickard, "Unilateral Markov fields,"Adv. Appl. Probab., vol. 12, pp. 655-671, 1980.
[30] D. K. Pickard, "Inference for discrete Markov fields: The simplest non-trivial case,"J. Amer. Statist. Assoc., vol. 82, pp. 90-96, 1982.
[31] A. Possolo, "Estimation of binary Markov random fields," Tech. Rep. 77, Dept. of Statistics, GN-22, Univ. of Washington, Seattle, WA, 1986.
[32] R. B. Potts, "Some generalised order-disorder transformations," inProc. Cambridge Phil. Soc., vol. 48, 1952, pp. 106-109.
[33] W. Qian and D. M. Titterington, "Pixel labeling for three-dimensional scenes based on Markov mesh models,"Signal Processing, vol. 22, pp. 313-328, 1991.
[34] W. Qian and D. M. Titterington, "Multidimensional Markov chain models for image textures,"J. Roy. Statist. Soc. B, vol. 53, pp. 661-674, 1991.
[35] W. Qian and D. M. Titterington, "Estimation of parameters in hidden Markov models,"Phil. Trans. Roy. Soc. Lond. A, vol. 337, pp. 407-428, 1991.
[36] B. D. Ripley and M. D. Kirkland, "Iterative simulation methods,"J. Comput. Appl. Math., vol. 31, pp. 165-172, 1990.
[37] B. Schachter and N. Ahuja, "Random pattern generation processes,"Computer Graphics Image Processing, vol. 10, pp. 95-114, 1979.
[38] R. H. Swendsen and J.-S. Wang, "Nonuniversal critical dynamics in Monte Carlo simulations,"Phys. Rev. Lett., vol. 58, pp. 86-88, 1987.
[39] A. M. W. Verhagen, "A three-parameter isotropic distribution of atoms and the hard-core square lattice gas,"J. Chem. Phys., vol. 67, pp. 5060-5065, 1977.
[40] C. N. Yang, "The spontaneous magnetization of a two-dimensional Ising model,"Phys. Rev., vol. 85, pp. 808-816, 1952.

Index Terms:
image reconstruction; Bayes methods; Markov processes; Markov random fields; Bayesian image restoration methods; moderate-to-large scale clustering; iterative algorithms; Markov mesh models; strong directional effects; parameter estimation
A.J. Gray, J.W. Kay, D.M. Titterington, "An Empirical Study of the Simulation of Various Models used for Images," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 16, no. 5, pp. 507-513, May 1994, doi:10.1109/34.291447
Usage of this product signifies your acceptance of the Terms of Use.