Using the Granold for Texture Classification

This paper describes the use the granold texture representation for image texture classification. The granold uses two different parameterised monotonic mappings to transform an input image into a function on two dimensions that may be regarded as a surface. The nature of this surface is such that corners appear at positions where there are simultaneously large changes in the response of the monotonic mappings to the input image. The shape and position of these corners is then analysed to provide information about the texture of the input image. Marginal probability mass functions are presented as a means to extract features from the granold for image texture classification purposes. A 16 class and a 2 class image texture classification experiment are described and their results discussed. The conclusion is that features extracted from the granold have discriminant power for image texture analysis.