Model-Based Pattern Recognition

  • Michal Haindl
Part of the Combinatorial Optimization book series (COOP, volume 13)


Recognition and processing of multi-dimensional data (or set of spatially related objects) is more accurate and efficient if we take into account all interdependencies between single objects. Objects to be processed like for example multi-spectral pixels in a digitized image, are often mutually dependent (e.g., correlated) with a dependency degree related to a distance between two objects in their corresponding data space. These relations can be incorporated into a pattern recognition process through appropriate multidimensional data model. If such a model is probabilistic we can use consistent Bayesian framework for solving many pattern recognition tasks. Data models are simultaneously useful to specify natural constraints and general assumptions about the physical world and a data capturing process hence they are essential in many data modelling or analytical procedures such as classification, segmentation, discontinuity detection, restoration, enhancement and scene analysis in general. Features derived from multi-dimensional data models are information preserving in the sense that they can be used to synthesize data spaces closely resembling original measurement data space as can be illustrated on the texture modelling application in the section 2.4.


Random Field Data Space Markov Random Field Multidimensional Data Texture Synthesis 
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.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    N. Ahuja and A. Rosenfeld, Mosaic models for textures, IEEE Trans. Pattern Anal. Mach. Int. Vol.3(1981) pp. 1–11.CrossRefGoogle Scholar
  2. [2]
    N. Ahuja and B.J. Schachter, Pattern models, (J.Wiley, 1983 ).Google Scholar
  3. [3]
    J. Bennett and A. Khotanzad, Maximum likelihood estimation methods for multispectral random field image models, IEEE Trans. on Pattern Analysis and Machine IntelligenceVol. 21 No. 3 (1999) pp. 537–543.CrossRefGoogle Scholar
  4. [4]
    M. Barnsley, Fractals Everywhere, (Academic Press, 1988 ).MATHGoogle Scholar
  5. [5]
    C. Bennis and A. Gagalowicz, Hierarchical texture synthesis on 3-D surfaces, in W. Hansmann, F.R.A. Hopgood, and W. Strasser (eds.) Eurographics ‘89, (1989) pp. 257–269.Google Scholar
  6. [6]
    C. Bennis and A. Gagalowicz, 2-D Macroscopic Texture Synthesis, Computer Graphics ForumVol. 8 (1989) pp. 291–300.CrossRefGoogle Scholar
  7. [7]
    J. Besag, Spatial Interaction and the Statistical Analysis of Lattice Systems, J. Royal Stat. Soc. Vol. B-36 (1974) pp. 192–236.MathSciNetGoogle Scholar
  8. [8]
    J. Besag, On the Statistical Analysis of Dirty Pictures, J. Royal Stat. Soc. Vol. B-48 (1986) pp. 259–302.MATHMathSciNetGoogle Scholar
  9. [9]
    C. Bouman and B. Liu, Multiple Resolution Segmentation of Textured Images, IEEE Trans. Pattern Anal. Mach. Int. Vol. 13 (1991) pp. 99–113.CrossRefGoogle Scholar
  10. [10]
    R. Chellappa, Two-dimensional discrete Gaussian Markov random field models for image processing, in L.N. Kanal and A. Rosenfeld (eds.) Progress in Pattern Recognition 2, ( Elsevier, North-Holland, 1985 ).Google Scholar
  11. [11]
    F.S. Cohen, Markov random fields for image modelling and analysis, in U.B. Desai (eds.) Modeling and Application of Stochastic Processes, ( Kluwer Academic Publishers, Boston, 1986 ).Google Scholar
  12. [12]
    G.R. Cross and A.K. Jain, Markov Random Field Texture Models, IEEE Trans. Pattern Anal. Mach. Int. Vol. 5 (1983) pp. 25–39.CrossRefGoogle Scholar
  13. [13]
    K. Deguchi and I. Morishita, Two-dimensional auto-regressive model for the representation of random image fields, Proc. 6-th Int. Conf. on Pattern Recognition, (Munich, 1982 ) pp. 90–93.Google Scholar
  14. [14]
    A. Dempster, N. Laird, and D. Rubin, Maximum likelihood from incomplete data via the EM algorithm, Journal of the Royal Statistical Society, B, Vol. 39 (1977) pp. 1–38.MATHMathSciNetGoogle Scholar
  15. [15]
    H. Derin and A.W. Cole, Segmentation of textured images using Gibbs random fields, Comp. Graphics Image Proc. Vol. 35 (1986) pp. 72–97.CrossRefGoogle Scholar
  16. [16]
    H. Del-in and H. Elliot, Modeling and segmentation of noisy and textured images using Gibbs random fields, IEEE Trans. Pattern Anal. Mach. Int. Vol. 9 (1987) pp. 39–55.CrossRefGoogle Scholar
  17. [17]
    N. Dodd, Multispectral Texture Synthesis Using Fractal Concepts, IEEE Trans. Pattern Anal. Mach. Int., Vol. 9 (1987) pp. 703–707.CrossRefGoogle Scholar
  18. [18]
    Z. Fan, F.S. Cohen, and M.A. Patel, Rotated and Scaled Textured Images using Markov Random Field Models, IEEE Trans. Pattern Anal. Mach. Int. Vol. 13 (1991) pp. 191–202.Google Scholar
  19. [19]
    O.D. Faugeras and O.D. Garber, Algebraic reconstruction techniques for texture synthesis, Proc. 5-th Int. Conf. on Pattern Recognition, (Miami Beach, 1980 ) pp. 792–782.Google Scholar
  20. [20]
    K.S. Fu and S.Y. Lu, Computer generation of texture using a syntactic approach, Computer Graphic, Vol. 12 (1978) pp. 147–152.Google Scholar
  21. [21]
    K. Fu, Syntactic Image Modelling Using Stochastic Tree Grammars, Comp. Graphics Image Proc. Vol. 12 (1980) pp. 136–152.CrossRefGoogle Scholar
  22. [22]
    S. Geman and D. Geman, Stochastic Relaxation, Gibbs Distributions and Bayesian Restoration of Images, IEEE Trans. Pattern Anal. Mach. Int., Vol. 6 (1984) pp. 721–741.CrossRefMATHGoogle Scholar
  23. [23]
    B. Gidas, A renormalization Group Approach to Image Processing problems, IEEE Trans. Pattern Anal. Mach. Int., Vol.11 (1989) pp.164–180.CrossRefMATHGoogle Scholar
  24. [24]
    J. Grim, On numerical evaluation of maximum likelihood estimates for finite mixtures of distributions, Kybernetika, Vol. 18 (1982) pp. 173–190.MATHMathSciNetGoogle Scholar
  25. [25]
    J. Grim and M. Haindl, Texture Modelling by Discrete Mixtures, accepted to Computational Statistics and Data Analysis, Elsevier (2002).Google Scholar
  26. [26]
    M. Haindl, Contextual Classification, Proc. Artificial intelligence application ‘80, (Prague, 1990 ).Google Scholar
  27. [27]
    M. Haindl, Texture Synthesis, CWI Quarterly, Vol. 4 (1991), pp. 305–331.MATHGoogle Scholar
  28. [28]
    M. Haindl and S. Šimberová, A Multispectral Image Line Reconstruction method, in P. Johansen and S. Olsen (eds.) Theory é4 Applications of Image Analysis, ( World Scientific Publishing Co., Singapore, 1992 ).Google Scholar
  29. [29]
    M. Haindl and S. Šimberová, A high-resolution radiospectrograph image reconstruction method, Astronomy & Astrophysics, Vol. 115 (1996) pp. 189–193.Google Scholar
  30. [30]
    M. Haindl and S. Šimberová, A Scratch Removal Method, Kybernetika, Vol. 34 (1998) pp. 423–428.Google Scholar
  31. [31]
    M. Haindl, Texture Segmentation Using Recursive Markov Random Field Parameter Estimation, in Bjarne K. Ersboll and Peter Johansen (eds.) Proceedings of the 11th Scandinavian Conference on Image Analysis, ( DSAGM, Kangerlussuaq, 1999 ) pp. 771–776.Google Scholar
  32. [32]
    M. Haindl, Recursive Model-Based Image Restoration, in A. Sanfeliu, J.J. Villanueva, M. Vanrell, R. Alquezar, T. Huang, and J. Serra (eds.) Proceedings of the 15th IAPR Int. Conf. on Pattern Recognition, ( Barcelona, IEEE Press, 2000 ) Vol. III, pp. 346–349.Google Scholar
  33. [33]
    M. Haindl, Texture modelling, in B. Sanchez, J. M. Pineda, J. Wolf-mann, Z. Bellahse, and F. Ferri (eds.) Proceedings of the World Multiconference on Systemics, Cybernetics and Informatics, (Orlando, International Institute of Informatics and Systemics, 2000 ), Vol. VII, pp. 634–639.Google Scholar
  34. [34]
    M. Haindl and V. Havlíček, A multiresolution causal colour texture model, in F. J. Ferri, J. M. Inesta, A. Amin, and P. Pudil (eds.) Advances in Pattern Recognition, Lecture Notes in Computer Science 1876, ( Berlin, Springer-Verlag, 2000 ) pp. 114–122.Google Scholar
  35. [35]
    J.M. Hammersley and D.C. Handscomb, Monte Carlo Methods, ( Methuen, London, 1964 ).CrossRefMATHGoogle Scholar
  36. [36]
    A. Jain, Advances in mathematical models for image processing, Proc. IEEE, Vol. 69 (1981) pp. 502–528.CrossRefGoogle Scholar
  37. [37]
    H. Kaneko and E. Yodogawa, A Markov Random Field Application to Texture Classification, Proc. IEEE Conf. Pattern Recognition and Image Processing, (Las Vegas, 1982 ).Google Scholar
  38. [38]
    R.L. Kashyap and R. Chellappa, Estimation and Choice of Neighbors in Spatial-Interaction Models of Images, IEEE Trans. Inf. Theory, Vol. 29 (1983) pp. 60–72.CrossRefMATHGoogle Scholar
  39. [39]
    R.L. Kashyap, Analysis and Synthesis of Image Patterns by Spatial Interaction Models, in L.N. Kanal and A. Rosenfeld (eds.) Progress in Pattern Recognition 1, ( Elsevier, North-Holland, 1981 ).Google Scholar
  40. [40]
    R. Kindermann and J.L. Snell, Markov Random Fields and their Applications, ( American Mathematical Society, Providence, 1980 ).Google Scholar
  41. [41]
    S.Y. Lu and K.S. Fu, A syntactic approach to texture analysis, Comp. Graphics Image Proc. Vol. 7 (1978) pp. 303–330.CrossRefGoogle Scholar
  42. [42]
    B.B. Mandelbrot, The Fractal Geometry of Nature, ( Freeman, San Francisco, 1982 ).MATHGoogle Scholar
  43. [43]
    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 (1953) pp. 1087–1091.CrossRefGoogle Scholar
  44. [44]
    S.G. Nadabar and A.K. Jain, Parameter Estimation in Markov Random Field Contextual Models Using Geometric Models of Objects, IEEE Trans. Pattern Anal. Mach. Int., Vol. 18 (1996) pp. 326–329.CrossRefGoogle Scholar
  45. [45]
    S. Peleg, J.Naor, R.Hartley, and D.Avnir, Multiple resolution texture analysis and classification, IEEE Trans. Pattern Anal. Mach. Int., Vol. 6 (1984) pp. 518–523.CrossRefGoogle Scholar
  46. [46]
    A.P. Pentland, Fractal-based description of natural scenes, IEEE Trans. Pattern Anal. Mach. Int., Vol. 6 (1984) pp. 661–674.CrossRefGoogle Scholar
  47. [47]
    D.K. Pickard, A curious binary lattice process, J. Appl.ProbabilityVol. 14 (1977) pp. 717–731.CrossRefMATHMathSciNetGoogle Scholar
  48. [48]
    Ch.J. Preston, Gibbs states on countable sets, (Cambridge Univ. Press, London, 1974 ).CrossRefMATHGoogle Scholar
  49. [49]
    W. Qian and D.M. Titterington, Multidimensional Markov Chain Models for Image Textures, J. Royal Stat. Soc., Vol.B-53 (1991) pp. 661–674.CrossRefGoogle Scholar
  50. [50]
    S. Richardson and P.J. Green, On Bayesian Analysis of Mixtures with and Unknown Number of Components, J. R. Statist. Soc., Vol. B-59 (1997) pp. 731–792.CrossRefMATHMathSciNetGoogle Scholar
  51. [51]
    S.J. Roberts, D. Husmeier, I. Rezek, and W. Penny, Bayesian Approaches to Gaussian Mixture Modeling, IEEE Trans. Pattern Anal. Mach. Int., Vol. 20 (1998) pp. 1133–1142.CrossRefGoogle Scholar
  52. [52]
    B.J. Schachter, A. Rosenfeld, and L.S. Davis, Random Mosaic Models for Textures, IEEE Trans. Syst. Man Cyb., Vol. 8 (1978) pp. 694–702.CrossRefGoogle Scholar
  53. [53]
    B.J. Schachter and N. Ahuja, Random pattern generation processing, Comp. Graphics Image Proc. Vol. 10 (1979) pp. 95–114.CrossRefGoogle Scholar
  54. [54]
    D. Stoyan, W.S. Kendall, and J. Mecke, Stochastic geometry and its applications, ( J.Wiley, 1989 ).Google Scholar
  55. [55]
    J.W. Woods, Two-Dimensional Discrete Markovian Fields, IEEE Trans. Inf. Theory, Vol. 18 (1972) pp. 232–240.CrossRefMATHMathSciNetGoogle Scholar
  56. [56]
    S.C. Zhu, X.W. Liu, and Y.N. Wu, Exploring Texture Ensembles by Efficient Markov Chain Monte Carlo-Toward a “Trichromacy” Theory of Texture, IEEE Trans. on Pattern Analysis and Machine IntelligenceVol. 22 (2000) pp. 554–569.CrossRefGoogle Scholar

Copyright information

© Kluwer Academic Publishers 2003

Authors and Affiliations

  • Michal Haindl
    • 1
  1. 1.Department of Pattern RecognitionInstitute of Information Theory and Automation, Academy of Sciences CRPragueCzech Republic

Personalised recommendations