A class of fractional random fields for image texture modelling

Instead of fractional Brownian motion, the authors consider the stochastic operational Laplace equation to model fractal characteristics of image data. Some properties, including the fractal dimension, of the random field generated from the Laplace equation are given. The spectral density of the random field is characterised by two parameters: a scaling exponent and a self-similarity index. A method is described to estimate these two parameters, which can be used to quantify and classify texture data.<>