A patchwork approach to stochastic simulation: A route towards the analysis of morphology in multiphase systems

摘要: Abstract We propose a new sequential stochastic simulation approach for black and white images in which we focus on the accurate reproduction of small scale geometry. Our aims at reproducing correctly connectivity properties geometry clusters are with respect to given length called block size. method is based analysis statistical relationships between adjacent square pieces image blocks. estimate transition probabilities blocks pixels training image. The simulations constructed by juxtaposing one pixels, hence term patchwork simulations. compare performance Strebelle’s multipoint algorithm several types increasing complexity. For composed size (e.g. squares, discs sticks), our produces better results than method. most noticeable improvement that cluster usually reproduced accurately. accuracy limited primarily Clusters significantly larger not As an example, applied this co-continuous polymer blend morphology as derived from electron microscope micrograph.