This article proposes the modelling and analysis of image texture using an extension of a locally stationary wavelet process model into two-dimensions for lattice processes. Such a model permits construction of estimates of a spatially localized spectrum and localized autocovariance which can be used to characterize texture in a multiscale and spatially adaptive way. We provide the necessary theoretical support to show that our two-dimensional extension is properly defined and has the proper statistical convergence properties. Our use of a statistical model permits us to identify, and correct for, a bias in established texture measures based on non-decimated wavelet techniques. The proposed method performs nearly as well as optimal Fourier techniques on stationary textures and outperforms them in non-stationary situations. We illustrate our techniques using pilled fabric data from a fabric care experiment and simulated tile data.

/pdf/locally-stationary-wavelet-fields-with-application-to-the-3ihxfei70i.pdf

Locally stationary wavelet fields with application to the modelling and analysis of image texture

A two-level system for the segmentation of texture images into regions of common textural properties is described. The first level is a purely numeric texture analyzer that uses texture energy measures to transform the image into feature measure planes. The second level is a knowledge-based segmentor that uses rules derived from knowledge of the image-forming process to arrive at a segmentation. Two different control schemes that can be used to guide the segmentation process are described. These are based on parallel region growing and iterative quadtree splitting, respectively. An illustration of the performance of the system with both control schemes on a real test image is presented

M. Simaan

Papers

Knowledge-guided segmentation of texture images