Parameter Estimation in a Gibbs-Markov Field Texture Model Based on a Coding Approach

Abstract

In this paper, we present a novel approach of the Conditional Least Square (CLS) estimator based on a coding scheme, for estimating the parameter vector associated with an Auto-Binomial model. This method provides a parallel solver for the estimation process. In order to illustrate the performance of the proposed approach, we carried out a Monte Carlo study and a real application for landscape classification using a high-resolution Pléiades-1A satellite image. Experimental results demonstrated the effectiveness of our estimation approach as well as CLS method, but in a lower runtime.