Lakhasly

Texture refers to properties that represent the surface or structure of an object (in re?ective or transmissive images, respectively); it is widely used, and perhaps intuitively obvious, but has no precise de?nition due to its wide variability.Texture consists of texture primitives or texture elements, sometimes called texels.The ?rst corresponds to textile checks or knitted stripes, and the second to the ?ner texture of the fabric or individual stitches.Texture is usually described as ?ne, coarse, grained, smooth, etc., implying that some more precise features must be de?ned to make machine recognition possible.A texture primitive is a contiguous set of pixels with some tonal and/or regional property, and can be described by its average intensity, maximum or minimum intensity, size, shape, etc.It is di?cult, however, to de?ne primitives for the checkered textile or fabric, which can be de?ned by at least two hierarchical levels.Examples are shown in Figure 15.1; dog fur, grass, river pebbles, cork, checkered textile, and knitted fabric.The main aim of texture analysis is recognition and texture-based shape analysis.

Texture refers to properties that represent the surface or structure of an object (in
reective or transmissive images, respectively); it is widely used, and perhaps intuitively
obvious, but has no precise denition due to its wide variability. We might dene texture
as something consisting of mutually related elements; therefore we are considering a
group of pixels and the texture described is highly dependent on the number considered
[Haralick, 1979]. Examples are shown in Figure 15.1; dog fur, grass, river pebbles, cork,
checkered textile, and knitted fabric.
Texture consists of texture primitives or texture elements, sometimes called texels. Primitives in grass and dog fur are represented by several pixels and correspond to
a stalk or a pile; cork is built from primitives that are comparable in size with pixels.
It is dicult, however, to dene primitives for the checkered textile or fabric, which can
be dened by at least two hierarchical levels. The rst corresponds to textile checks or
knitted stripes, and the second to the ner texture of the fabric or individual stitches.
As we have seen in many other areas, this is a problem of scale; texture description is
scale dependent.
The main aim of texture analysis is recognition and texture-based shape analysis.
Textured properties of regions were referred to many times while considering image segmentation (Chapter 6), and derivation of shape from texture was discussed in Chapter 11. Texture is usually described as ne, coarse, grained, smooth, etc., implying
that some more precise features must be dened to make machine recognition possible.
Such features can be found in the texture tone and structure [Haralick, 1979]. Tone is
based mostly on pixel intensity properties in the primitive, while structure is the spatial
relationship between them.
Each pixel can be characterized by its location and tonal properties. A texture
primitive is a contiguous set of pixels with some tonal and/or regional property, and
can be described by its average intensity, maximum or minimum intensity, size, shape,
etc. The spatial relationship of primitives can be random, or they may be pairwise
dependent, or some number of primitives can be mutually dependent. Image texture is
then described by the number and types of primitives and by their spatial relationship.