Noncausal Estimation for Discrete Gauss-Markov Random Fields
In , it was shown that 2-D discrete Gauss-Markov random fields can be characterized in terms of a noncausal nearest-neighbor model (NNM) driven by locally correlated noise. This result is used here to obtain a simple solution of the smoothing problem for Gauss-Markov random fields. It is shown that the smoother has a nearest-neighbor structure of the same type as the original field, and that the smoothing error is itself a Gauss-Markov random field. Since the operator describing the smoother dynamics is positive and self-adjoint, the smoother can be implemented by using efficient iterative algorithms for elliptic PDEs.
KeywordsImage Restoration Gibbs Distribution Elliptic PDEs Boundary Process Reciprocal Process
Unable to display preview. Download preview PDF.
- J. Besag, “On the statistical analysis of dirty pictures (with discussion),” J. Royal Statistical Soc., series B, vol. 48, no. 3, pp. 259–302, 1986.Google Scholar
- F. S. Cohen and D. B. Cooper, “Simple parallel hierarchical and relaxation algorithms for segmenting noncausal Markovian random fields,” IEEE Trans. Pattern Analysis, Machine Intell., vol. PAMI-9, no. 2, March 1987.Google Scholar
- J. Besag, “Spatial interaction and the statistical analysis of lattice systems (with discussion),” J. Royal Statistical Soc., series B, vol. 36, pp. 192–236, 1974.Google Scholar
- M. S. Bartlett, The Statistical Analysis of Spatial Pattern, Chapman and Hall, London, 1975.Google Scholar
- A. J. Krener, “Reciprocal diffusions and stochastic differential equations of second order,” Stochastics, vol. 24, pp. 393–422, 1988.Google Scholar
- B. C. Levy, R. Frezza, and A. J. Krener, “Modeling and estimation of discretetime Gaussian reciprocal processes,” technical report, Dept. of Elec. Eng. and Comp. Science, UC Davis, CA, March 1989.Google Scholar
- M. B. Adams, “Linear estimation of boundary value stochastic processes,” Ph. D. thesis, Dept. of Aeronautics and Astronautics, MIT, Cambridge, MA, Jan. 1983.Google Scholar
- B. C. Levy, M. B. Adams, and A. S. Willsky, “Solution and linear estimation of 2-D nearest-neighbor models,” to appear in Proc. IEEE (special issue on multidimensional signal processing), May 1990.Google Scholar
- D. M. Young, Iterative Solution of Large Linear Systems, Academic Press, New York, 1971.Google Scholar
- P. Concus, G. H. Golub, and D. P. O’Leary, “A generalized conjugate gradient method for the numerical solution of elliptic partial differential equations,” in Sparse Matrix Computations, J.R. Bunch and D. J. Rose eds., Academic Press, New York, pp. 309–332.Google Scholar
- S. McCormick, Multigrid Methods, SIAM, Philadelphia, PA, 1987.Google Scholar
- R. Glowinski, G. H. Golub, G. A. Meurant, and J. Périaux, eds., First International Symposium on Domain Decomposition Methods for Partial Differential Equations, SIAM, Philadelphia, PA, 1988.Google Scholar