<title>SAR image segmentation using generalized pairwise Markov chains</title>

Abstract

The efficiency of Markov models in the context of SAR image segmentation mainly relies on their spatial regularity constraint. However, a pixel may have a rather different visual aspect when it is located near a boundary or inside a large set of pixels of the same class. According to the classical hypothesis in Hidden Markov Chain (HMC) models, this fact can not be taken into consideration. This is the very reason of the recent Pairwise Markov Chains (PMC) model which relies on the hypothesis that the pairwise process (X, Y ) is Markovian and stationary, but not necessarily X. The main interest of the PMC model in SAR image segmentation is to not assume that the speckle is spatially uncorrelated. Hence, it is possible to take into account the difference between two successive pixels that belong to the same region or that overlap a boundary. Both PMC and HMC parameters are learnt from a variant of the Iterative Conditional Estimation method. This allows to apply the Bayesian Maximum Posterior Marginal criterion for the restoration of X in an unsupervised manner. We will compare the PMC model with respect to the HMC one for the unsupervised segmentation of SAR images, for both Gaussian distributions and Pearson system of distributions.

Variational Bayesian Inference for Pairwise Markov Models

“…In this paper, we start by including all these models into a common probabilistic model called the Pairwise Markov Model (PMM) [13]. This model has been introduced in a Bayesian framework where the objective is to estimate the latent process from the observed one [14,15,16,17]. Next, i) we focus on the generative aspect of the PMM, and show that this model p θ (.)…”

“…This results generalizes the form of the previous covariance matrices as it can be checked by setting e = f = 0 or e = 0, for example. At this point of our work, we have not identified if the full linear and Gaussian PMM p θ (x) can model any Gaussian distribution with a covariance matrix satisfying (15), except in some particular cases (see e.g. Proposition 1).…”

Independent feature subspace iterative optimization based fuzzy clustering for synthetic aperture radar image segmentation

“…Researchers have proposed many kinds of segmentation algorithms for SAR images, which include threshold methods, 3,[17][18][19] spectral clustering (SC) algorithms, 20,21 statistic model-based methods, 1,14,15,[22][23][24][25] artificial intelligence methods, [26][27][28][29][30] support vector machine (SVM), 6,31 region growing methods, 15,[32][33][34][35] and so on. Among these algorithms, cluster-based algorithms form one popular and representative family, whose main idea is to group pixels in such a way that the pixels in the same group are more similar to each other than those in other groups.…”

Particle filtering with pairwise Markov processes

“…Propagatingp(x, lygtn) amounts to propagating these parameters, and in this case equation (8) reduces to the Kalman filter. However, in the general case, computing equation (8) is difficult in practice. Consequently a number of approximate, Monte Carlo based methods have been derived.…”