Multidimensional Markov Chain Models for Image Textures
Wenbin Qian, D. Michael Titterington · Journal of the Royal Statistical Society Series B (Statistical Methodology) · 1991
SUMMARY Multidimensional Markov chain models are developed for texture. By suitable choice of parameters, textures can be simulated that are similar to those generated by Markov random field models, but the simulation procedure is computationally much more economical. The problem of parameter estimation is examined for non-noisy and noisy data. In the latter case, procedures are developed for simultaneous parameter estimation and image restoration. Illustrative examples are provided.