DEFINITION AND APPLICATION OF A FOURIER DOMAIN TEXTURE MEASURE - APPLICATIONS TO HISTOLOGICAL IMAGE SEGMENTATION

A texture measure is defined for the purpose of two-dimensional histol ogical image classification, based upon a novel exploitation of the mo dulus of the Fourier transform. Its implementation in the form of an a lgorithm is presented and discussed, and employed to classify a set of histological images. The results, which were obtained by the analysis and classification of different mammalian tissue types, compare favou rably with established texture recognition techniques. Furthermore, th e results suggest that the method has a comparable performance to othe r techniques at reduced storage and time costs for images of a high gr ey level resolution.