Restriction of a Markov random field on a graph and multiresolution statistical image modeling

The association of statistical models and multiresolution data analysis in a consistent and tractable mathematical framework remains an intricate theoretical and practical issue. Several consistent approaches have been proposed previously to combine Markov random field (MRF) models and multiresolution algorithms in image analysis: renormalization group, subsampling of stochastic processes, MRFs defined on trees or pyramids, etc. For the simulation or a practical use of these models in statistical estimation, an important issue is the preservation of the local Markovian property of the representation at the different resolution levels. It is shown that this key problem may be studied by considering the restriction of a Markov random field (defined on some simple finite nondirected graph) to a part of its original site set. Several general properties of the restricted field are derived. The general form of the distribution of the restriction is given. "Locality" of the field is studied by exhibiting a neighborhood structure with respect to which the restricted field is an MRF. Sufficient conditions for the new neighborhood structure to be "minimal" are derived. Several consequences of these general results related to various "multiresolution" MRF-based modeling approaches in image analysis are presented.

[1]  Maria Petrou,et al.  A generalisation of renormalisation group methods for multiresolution image analysis , 1992, [1992] Proceedings. 11th IAPR International Conference on Pattern Recognition.

[2]  Josef Kittler,et al.  Multiresolution motion segmentation , 1994, Proceedings of 12th International Conference on Pattern Recognition.

[3]  Robert Azencott,et al.  Simulated annealing : parallelization techniques , 1992 .

[4]  Jun Zhang,et al.  A Markov Random Field Model-Based Approach to Image Interpretation , 1992, IEEE Trans. Pattern Anal. Mach. Intell..

[5]  Fabrice Heitz,et al.  Hierarchical statistical models for the fusion of multiresolution image data , 1995, Optics + Photonics.

[6]  Mark R. Luettgen,et al.  Image processing with multiscale stochastic models , 1993 .

[7]  Hans-Otto Georgii,et al.  Gibbs Measures and Phase Transitions , 1988 .

[8]  Eric Dubois,et al.  Bayesian Estimation of Motion Vector Fields , 1992, IEEE Trans. Pattern Anal. Mach. Intell..

[9]  J. Laurie Snell,et al.  Markov Random Fields and Their Applications , 1980 .

[10]  Charles A. Bouman,et al.  A multiscale random field model for Bayesian image segmentation , 1994, IEEE Trans. Image Process..

[11]  W. Clem Karl,et al.  Efficient multiscale regularization with applications to the computation of optical flow , 1994, IEEE Trans. Image Process..

[12]  J. Besag.  Spatial Interaction and the Statistical Analysis of Lattice Systems , 1974 .

[13]  Davi Geiger,et al.  Scaling images and image features via the renormalization group , 1993, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.

[14]  Josiane Zerubia,et al.  Multiscale Markov random field models for parallel image classification , 1993, 1993 (4th) International Conference on Computer Vision.

[15]  A. Krener,et al.  Modeling and estimation of discrete-time Gaussian reciprocal processes , 1990 .

[16]  Xavier Descombes.  Champs markoviens en analyse d'images , 1993 .

[17]  Patrick Bouthemy,et al.  Global Bayesian Estimation, Contrained Multiscale Markov Random Fields and the Analysis of Visual Motion , 1993 .

[18]  François Charot,et al.  Efficient Parallel Nonlinear Multigrid Algorithms for Low-Level Vision Applications , 1995, J. Parallel Distributed Comput..

[19]  Patrick Bouthemy,et al.  Multimodal Estimation of Discontinuous Optical Flow using Markov Random Fields , 1993, IEEE Trans. Pattern Anal. Mach. Intell..

[20]  Francoise J. Preteux,et al.  Hierarchical Markov random field models applied to image analysis: a review , 1995, Optics + Photonics.

[21]  J. Davenport.  Editor , 1960 .

[22]  M. Hurn,et al.  A study of simulated annealing and a revised cascade algorithm for image reconstruction , 1995 .

[23]  P. Pérez,et al.  Multiscale minimization of global energy functions in some visual recovery problems , 1994 .

[24]  Tomaso Poggio,et al.  Probabilistic Solution of Ill-Posed Problems in Computational Vision , 1987 .

[25]  I. Matsuba.  Renormalization group approach to hierarchical image analysis , 1988, ICASSP-88., International Conference on Acoustics, Speech, and Signal Processing.

[26]  Fabrice Heitz,et al.  Restriction of a Markov random field on a graph and multiresolution image analysis , 1994 .

[27]  Demetri Terzopoulos,et al.  Image Analysis Using Multigrid Relaxation Methods , 1986, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[28]  Basilis Gidas,et al.  A Renormalization Group Approach to Image Processing Problems , 1989, IEEE Trans. Pattern Anal. Mach. Intell..

[29]  W. C. Karl,et al.  Eecient Multiscale Regularization with Applications to the Computation of Optical Flow 1 , 1993 .

[30]  J. Besag.  On the Statistical Analysis of Dirty Pictures , 1986 .

[31]  J. M. Dinten,et al.  Tomographic reconstruction of axially symmetric objects: Regularization by a Markovian modelization , 1990, [1990] Proceedings. 10th International Conference on Pattern Recognition.

[32]  Donald Geman,et al.  Boundary Detection by Constrained Optimization , 1990, IEEE Trans. Pattern Anal. Mach. Intell..

[33]  Charles A. Bouman,et al.  Multiple Resolution Segmentation of Textured Images , 1991, IEEE Trans. Pattern Anal. Mach. Intell..

[34]  Furs-Ching Jeng,et al.  Subsampling of Markov random fields , 1992, J. Vis. Commun. Image Represent..

[35]  Michèle Basseville,et al.  Modeling and estimation of multiresolution stochastic processes , 1992, IEEE Trans. Inf. Theory.

[36]  Donald Geman,et al.  Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images , 1984, IEEE Transactions on Pattern Analysis and Machine Intelligence.