A Multiscale Test of Spatial Stationarity for Textured Images in R

The ability to automatically identify areas of homogeneous texture present within a greyscale image is an important feature of image processing algorithms. This article describes the R package LS2Wstat which employs a recent wavelet-based test of stationarity for locally stationary random fields to assess such spatial homogeneity. By embedding this test within a quadtree image segmentation procedure we are also able to identify texture regions within an image.