ABSTRACT
Texture is the quality of region of an image that can include smoothness, regularity of a pattern, or variations in the pattern. Texture for a single pixel requires information from the surrounding pixels and therefore is a higher-ordered calculation. This chapter reviews some of the methods and organize these methods into classes according to their principles. The Brodatz data set is a long standing set of texture-based images. This data set is available through many different websites. These are quite large, and commonly, researchers will cut these large images into smaller subimages. Decomposition of an image using wavelets is performed through iterations of downsampling and edge enhancement. The decomposition of the image is an iterative process, in which each iteration downsamples the image, creates three edge-enhanced images, and then places these four images into the larger frame. The decomposition image contains edge information across a wide range of scales.
