ABSTRACT

The English word texture has a common etymology with the word textile. Within the context of visual perception, the term texture refers to surface appearance characterized by a regular pattern distributed over the surface. Figures 8.1 and 8.2 show examples of visual textures, both artificial and naturally occurring. The regularity associated with a visual texture can be highly structured, as in a brick wall, or more random, as in the view of a grassy field. Visual textures play several important roles in perception (Landy & Graham, 2004). Most obviously, recognizable textures can indicate the materials composing an object’s surfaces. Even when visual textures cannot be identified as a specific material, they may provide information about the material properties of a surface. Surfaces in the world are often made of different materials, and so visual texture can often be used to organize a two-dimensional view of the world into different regions likely to correspond to different objects or object parts. Finally, variability in visual textures over the two-dimensional, projected view of the world often provides evidence for aspects of the three-dimensional geometric structure of the

To exploit the information available in visual textures, the perceptual system must extract characterizations of the optical variability of the textures that correlate with material and geometric properties of the environment under view. This is a problem that has been studied extensively in both the vision science and computer vision communities. Historically, two ways of characterizing visual textures were common: structural and statistical.