Estimation of Markov Random Field Parameters Using Ant Colony Optimization for Continuous Domains

In this paper, we present a method based on ant colony optimization for continuous domains (ACOC) to estimate the Markov random field parameters, using the maximum likelihood criterion. In order to model the multi-level image patterns more accurately, we define a new clique potential function. Experimental results and performance comparison with the Markov chain Monte Carlo method are provided to illustrate the performance of the ACOC-based method.