Analysis of Texture Anisotropy Based on Some Gaussian Fields with Spectral Density

International audienceIn this paper, we describe a statistical framework for the analysis ofanisotropy of image texture. This framework is based on the modeling of the imageby two kinds of non-stationary anisotropicGaussian field with stationary incrementsand spectral density: the extended fractional Brownian field (EFBF) and a specificGaussian operator scaling field (GOSF), which both correspond to a generalizationof the fractional Brownian field. In this framework, we tackle anisotropy analysisusing some directional processes that are either defined as a restriction of the imageon an oriented line or as a projection of the image along a direction. In the context ofEFBF and GOSF, we specify links between the regularity of line and projection processesand model parameters, and explain how field anisotropy can be apprehendedfrom the analysis of process regularity. Adapting generalized quadratic variations,we also define some estimators of the regularity of line and projection processes,and study their convergence to field model parameters. Estimators are also evaluatedon simulated data, and applied for illustration to medical images of the breastand the bone