Stochastic models in materials science

Improving the understanding of the link between materials' physical properties and their microstructures is one of the objectives of materials science. Stochastic models can be exploited to investigate this relation. Microstructures are observed on three-dimensional images obtained by micro-computed tomography. Once the geometric features are extracted on basis of image processing, it is possible to construct a model that embeds the observed characteristics' distributions. We introduce some stochastic models for fiber systems and foams. Our focus is the investigation of which characteristics are crucial for model fitting. Strategies to facilitate the choice of the parameters of the model are discussed.