The term texture is a somewhat confusing term in computer graphics and generally does not mean controlling the small-scale geometry of the surface of a computer graphics object, which is the normal meaning of the word. A better term is colour map. Texture is mapped onto an already available surface. A surface texture is created by the regular repetition of an element or pattern, called surface texel, on a surface. In computer graphics there are deterministic (regular) and statistical (irregular) textures. Deterministic texture is created by repetition of a fixed geometric shape such as a circle or square. Examples of deterministic textures are patterned wallpaper and bricks. Texels are represented naturally by the shape parameters of the specific shape. Statistical textures are created by changing patterns with fixed statistical properties. Most natural textures like wood or stone are statistical. Statistical textures are represented typically in term of spatial frequency properties.

Texture and texture mapping became a highly developed tool in the 1980s and is the technique used to enhance scenes so that they are more visually interesting. Its main attraction is cheapness. It can be grafted onto a standard rendering method without adding too much to the processing cost. The mapping of texture has become consolidated with the advent of cheap texture mapping hardware facilities. This kind of hardware is able to map a 2D texture onto a surface that is represented by a polygon mesh.

Mapping a 2D texture map onto the surface of an object then projecting the object into screen space is a 2D to 2D transformation and can be viewed as an image warping operation. The most common way to do this is to inverse map. For each pixel we find its corresponding pre-image in texture space. The pre-image is the area in texture space that is projected onto one pixel on the surface. However, specifying this overall transformation is not straightforward. Normally, texture mapping is a two-stage process that takes us from the 2D texture space into the 3D space of the object and then via the projective transform into 2D screen space. The first transformation is known as parametrisation and the second stage is the normal computer graphics projective transformation. Parametrisation associates all points in texture space with points on the object surface.

The problem with using texture is that aliasing problems in texture mapping are usually highly visible. By definition texture usually manifests some kind of coherence or periodicity. The aliasing effect occurs as the periodicity in the texture approaches the pixel resolution. Therefore, the use of an anti-aliasing method is mandatory with texture mapping. This is easily seen by considering an object retreating away from a viewer so that its projection in screen space covers fewer and fewer pixels. As the object size decreases, the pre-image of a pixel in texture space will increase covering a larger area. If we simply point sample at the centre of the pixel and take the value at the corresponding point in texture space, then incorrect results will follow. Anti-aliasing means integrating the information over the pixel pre-image. At best we can approximate this integral because we have no knowledge of the shape of the quadrilateral arising from the pre-image of the pixel, only its four corner points.