TEXTURE RECONSTRUCTION IN NOISY IMAGES

An approach for texture reconstruction in noisy images is presented. The reconstruction Is achieved by a denoising algorithm using Gauss Markov random fields as texture models and taking advantage of both the first and the second level of Bayesian inference to obtain a MAP estimate of the noise-free image and an estimate of the model parameters, respectively. For noise-free images exist several techniques for parameter estimation. In the noisy case, however, parameter estimation becomes much more difficult. In order to obtain the best possible texture reconstruction we propose an iterative algorithm to estimate the parameters that provide the highest evidence. This results in an optimization problem where we use the current MAP estimate of the noise-free image to calculate an approximated model evidence for a given set of parameters and adjust these parameters accordingly.

[1]  Gerhard Winkler,et al.  Image analysis, random fields and dynamic Monte Carlo methods: a mathematical introduction , 1995, Applications of mathematics.

[2]  John Skilling,et al.  Data analysis : a Bayesian tutorial , 1996 .

[3]  Richard E. Blahut,et al.  Principles and practice of information theory , 1987 .

[4]  Mihai Datcu,et al.  Spatial information retrieval from remote-sensing images. I. Information theoretical perspective , 1998, IEEE Trans. Geosci. Remote. Sens..

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

[6]  Sridhar Lakshmanan,et al.  Simultaneous Parameter Estimation and Segmentation of Gibbs Random Fields Using Simulated Annealing , 1989, IEEE Trans. Pattern Anal. Mach. Intell..

[7]  Haluk Derin,et al.  Modeling and Segmentation of Noisy and Textured Images Using Gibbs Random Fields , 1987, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[8]  Christopher J. Oliver,et al.  Optimum texture analysis of SAR images , 1994, Defense, Security, and Sensing.

[9]  D. Geman.  Random fields and inverse problems in imaging , 1990 .

[10]  Anil K. Jain,et al.  Texture fusion and feature selection applied to SAR imagery , 1997, IEEE Trans. Geosci. Remote. Sens..

[11]  W. von der Linden,et al.  The Prior-Predictive Value: A Paradigm of Nasty Multi-Dimensional Integrals , 1999 .

[12]  C. Oliver.  Information from SAR images , 1991 .

[13]  David J. C. MacKay,et al.  Bayesian Interpolation , 1992, Neural Computation.

[14]  Rama Chellappa,et al.  Unsupervised Texture Segmentation Using Markov Random Field Models , 1991, IEEE Trans. Pattern Anal. Mach. Intell..

[15]  J. Goodman.  Statistical Properties of Laser Speckle Patterns , 1963 .