This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
A Markov Random Field Model-Based Approach to Image Interpretation
June 1992 (vol. 14 no. 6)
pp. 606-615

An image is segmented into a collection of disjoint regions that form the nodes of an adjacency graph, and image interpretation is achieved through assigning object labels (or interpretations) to the segmented regions (or nodes) using domain knowledge, extracted feature measurements, and spatial relationships between the various regions. The interpretation labels are modeled as a Markov random field (MRF) on the corresponding adjacency graph, and the image interpretation problem is then formulated as a maximum a posteriori (MAP) estimation rule, given domain knowledge and region-based measurements. Simulated annealing is used to find this best realization or optimal MAP interpretation. This approach also provides a systematic method for organizing and representing domain knowledge through appropriate design of the clique functions describing the Gibbs distribution representing the pdf of the underlying MRF. A general methodology is provided for the design of the clique functions. Results of image interpretation experiments on synthetic and real-world images are described.

[1] T. E. Avery,Interpretation of Aerial Photographs. Minneapolis: Burgess, 1977.
[2] A. Rosenfeld and A. Kak,Digital Picture Processing, New York: Academic, 1976.
[3] B. K. P. Horn,Robot Vision. Cambridge, MA: M.I.T. Press, 1986.
[4] R. A. Schowengerdt,Techniques for Image Processing and Classification in Remote Sensing. New York: Academic, 1983.
[5] R. D. Keller,Expert System Technology: Development and Application. Englewood Cliffs, NJ: Lourdon, 1987.
[6] M. Nagao and T. Matsuyama,A Structural Analysis of Complex Aerial Photographs. New York: Plenum, 1980.
[7] T. Binford, "Survey of model-based image analysis systems,"Int. J. Robotics Res., vol. 1, no. 1, pp. 587-633, 1982.
[8] Y. Ohta,Knowledge-Based Interpretation of Outdoor Natural Color Scenes. Boston: Pitman, 1985.
[9] O. D. Faugeras and K. E. Price, "Semantic description of aerial images using stochastic relaxation,"IEEE Trans. Patt. Anal. Machine Intell., vol. PAMI-3, pp. 633-642, Nov. 1981.
[10] R. Prasannappa, L. Davis, and V. S. S. Hwang, "A knowledge-based vision system for aerial image understanding," CS-TR-1758, Univ. of Maryland, Jan. 1987.
[11] R. Brooks, "Symbolic reasoning among 3-dimensional models and 2- dimensional images,"Artificial Intelligence, vol. 17, pp. 285-394, 1981.
[12] D. M. McKeown, W. A. Harvey, and J. McDermott, "Rule-based interpretation of aerial imagery,"IEEE Trans. Patt. Anal. Machine Intell., vol. PAMI-7, pp. 570-585, Sept. 1985.
[13] B. Draper, R. Collins, J. Brolio, A. R. Hanson, and E. M. Riseman, "The schema system," Tech. Rep., COINS 88-76, Univ. of Massachusetts, Sept. 1988.
[14] O. Frank and D. Strauss, "Markov graphs,"J. Amer. Stat. Assoc., vol. 81, pp. 832-842, Sept. 1986.
[15] G. C. Cross and A. K. Jain, "Markov random field texture models,"IEEE Trans. Patt. Anal. Machine Intell., vol. PAMI-5, pp. 25-39, 1983.
[16] S. Geman and D. Geman, "Stochastic relaxation, Gibbs distribution and the Bayesian restoration of images,"IEEE Trans. Patt. Anal. Machine Intell., vol. PAMI-6, pp. 721-741, Nov. 1984.
[17] C. W. Therrien, T. F. Quatieri, and D. E. Dudgeon, "Statistical model-based algorithms for image analysis,"Proc. IEEE, vol. 74, Apr. 1986.
[18] J. Zhang and J. W. Modestino, "Markov random fields with applications to texture classification and discrimination," inProc. 20th Ann. Conf. Inform. Sci. Syst. (Princeton, NJ), Mar. 1986.
[19] 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.
[20] F. S. Cohen and D. B. Cooper, "Simple parallel hierarchical and relaxation algorithms for segmenting noncausal Markovian fields,"IEEE Trans. Pattern Anal. Machine Intell., vol. PAMI-9, pp. 195-219, March 1987.
[21] J. Besag, "On the statistical analysis of dirty pictures,"J. Royal Stat. Soc. B, vol. 48, pp. 259-302, 1986.
[22] C. Bouman and B. Liu, "Segmentation of textured images using a multiple resolution approach," inProc. IEEE Int. Conf. Acoust., Speech, Signal Processing, New York, Apr. 11-14, 1988, pp. 1124-1127.
[23] J. Marroquin, S. Mitter, and T. Poggio, "Computer vision,"J. Amer. Stat. Assoc., vol. 82, pp. 76-89, Mar. 1987.
[24] P. B. Chou and C. M. Brown, "Multi-modal segmentation using Markov random fields," inProc. IJCAI, 1987, pp. 663-670.
[25] R. Kinderman and J. L. Snell,Markov Random Fields and Their Applications. Providence, RI: Amer. Math. Soc., 1980.
[26] J. Besag, "Spatial interaction and the statistical analysis of lattice systems,"J. Roy. Stat. Soc. Series B, vol. 36, pp. 192-226, 1974.
[27] J. A. Feldman and Y. Yakimovsky, "Decision theory and artificial intelligence: I. A semantic-based region analyzer,"Artificial Intell., vol. 5, pp. 325-348, 1974.
[28] J. W. Modestino, "A hierarchical region-based approach to automated photointerpretation," RPI Rep., Mar. 1987.
[29] L. Zadeh, "Approximate reasoning based on fuzzy logic," inProc. 6th IJCAI, 1979, pp. 1004-1010.
[30] S. Kirkpatrick, C. S. Gelatt, and M. P. Vecchi, "Optimization by simulated annealing,"Sci., vol. 220, pp. 671-680, May 1983.
[31] A. A. El Gamal, L. A. Hemachandra, I. Shperling, and V. K. Wei, "Using simulated annealing to design good codes,"IEEE Trans. Inform. Theory, vol. IT-33, no. 1, pp. 116-123, Jan. 1987.
[32] J. Zhang, "Two-dimensional stochastic model-based image analysis," Ph.D. thesis, Rensselaer Polytechnic Inst., Troy, NY, Aug. 1988.
[33] J. Zhang and J. W. Modestino, "Image segmentation using a Gaussian model," inProc. Conf. Inform. Sci. Syst.(Princeton, NJ), Mar. 1988.
[34] J. Kanai, "Interpretation of real images using the MRF model-based method," RPI Rep., Sept. 1987.
[35] L. Y. Yu, "New results on image segmentation and interpretation using the MRF model," RPI Rep., June 1988.
[36] A. Rosenfeld, R. A. Hummel, and S. W. Zucker, "Scene labeling by relaxation operations,"IEEE Trans. Syst. Man Cybern., vol. SMC-6, pp. 420-433, 1976.
[37] R. M. Haralick and L. G. Shapiro, "The consistent labeling problem: Part I,"IEEE Trans. Patt. Anal. Machine Intell., vol. PAMI-1, pp. 173-184, Apr. 1979.
[38] R. M. Haralick and L. G. Shapiro, "The consistent labeling problem: Part II,"IEEE Trans. Patt. Anal. Machine Intell., vol. PAMI-2, pp. 193-203, May 1980.
[39] R. A. Hummel and S. W. Zucker, "On the foundation of relaxation labeling process,"IEEE Trans. Patt. Anal. Machine Intell., vol. PAMI-5, pp. 267-287, May 1983.

Index Terms:
picture processing; pattern recognition; simulated annealing; synthetic images; Markov random field model-based approach; image interpretation; disjoint regions; adjacency graph; object labels; domain knowledge; extracted feature measurements; spatial relationships; interpretation labels; real-world images; graph theory; Markov processes; pattern recognition; picture processing; simulated annealing
Citation:
J.W. Modestino, J. Zhang, "A Markov Random Field Model-Based Approach to Image Interpretation," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 14, no. 6, pp. 606-615, June 1992, doi:10.1109/34.141552
Usage of this product signifies your acceptance of the Terms of Use.