Concrete CT Image Segmentation Using Modified Metropolis Dynamics
Liang Zhao, Changhua Li, Dengfeng Chen, Faning Dang · 2009
In this paper, we present a pseudo-stochastic variation of the Metropolis dynamics for combinatorial optimization in concrete CT image classification using Markov Random Fields. The method is a modified version of the Metropolis (MMD) algorithm: at each iteration, the new state is chosen randomly, but the decision to accept it is purely deterministic. This is also a suboptimal technique but it is much faster than stochastic relaxation. Experimental results are compared to those obtained by the Metropolis algorithm, the Gibbs sampler and ICM (Iterated Conditional Mode). Classify result indicate that using MMD can reflect the spatial distribution of the concrete materials on deformation, and afford an effective method on concrete meso-structure computerized tomography (CT) image study.