Texture

Overview

Texture in everyday language is a property of surfaces associated with the tactile quality they suggest (texture has the same root as textile). In computational vision, it refers to a closely related concept, that of a spatially repeating pattern on a surface that can be sensed visually. Examples include the pattern of windows on a building, the stitches on a sweater, the spots on a leopard's skin, blades of grass on a lawn, pebbles on a beach or a crowd of people in a stadium. Sometimes the arrangement is quite periodic, as in the stitches on a sweater; in other instances, as in pebbles on a beach, the regularity is only statistical--the density of pebbles is roughly the same on different parts of the beach.

Texture can play several roles in visual perception:

Texture segmentation can be used to find the boundary contours of the objects. These contours can be used to fit volumetric primitives, compute invariants, or match with 2D aspects, as done in currently popular object recognition approaches with brightness edges.
Texture in itself provides a cue for recognition. Think of the spots on a giraffe, or the stripes on a zebra. To exploit this cue, one should be able to construct texture descriptors that are stable under changes of viewpoint and lighting.
Shape from texture analysis could be used to recover geometrical properties for indexing and matching with object models. Shape here is intended to have a rather broad meaning--it is not necessarily a dense surface orientation map
Texture synthesis is one application of texture analysis in computer graphics. The goal is to synthesize novel samples of a given texture, which, while not identical, will be percieved by humans to be the same texture.

Publications

A. A. Efros & W. T. Freeman, "Image Quilting for Texture Synthesis and Transfer", Proceedings of SIGGRAPH '01, 2001. [pdf] [more info]

D. Forsyth, "Shape from texture and integrability" International Conference on Computer Vision, Vancouver, BC, 2001. [pdf] [ps]

T. K. Leung & J. Malik, "Representing and Recognizing the Visual Appearance of Materials using Three-dimensional Textons", International Journal of Computer Vision, 2000. [pdf] [ps] [more info]

A. A. Efros & T. K. Leung, "Texture Synthesis by Non-Parametric Sampling", International Conference on Computer Vision, Corfu, Greece, 1999. [pdf] [ps] [more info]

T. K. Leung & J. Malik, "On Perpendicular Texture: Why do we see more flowers in the distance?", Conf. Computer Vision and Pattern Recognition, 1997 [pdf] [ps] [more info]

T. K. Leung & J. Malik, "Detecting, Localizing and Grouping Repeated Scene Elements from an Image", European Conference on Computer Vision, Cambridge, England, 1996 [pdf] [ps] [more info]

J. Malik & R. Rosenholtz, "Computing local surface orientation and shape from texture for curved surfaces." International Journal of Computer Vision 23(2):149-168. [pdf] [ps]

Back to Berkeley Computer Vision page

Publications
	A. A. Efros & W. T. Freeman, "Image Quilting for Texture Synthesis and Transfer", Proceedings of SIGGRAPH '01, 2001. [pdf] [more info]
	D. Forsyth, "Shape from texture and integrability" International Conference on Computer Vision, Vancouver, BC, 2001. [pdf] [ps]
	T. K. Leung & J. Malik, "Representing and Recognizing the Visual Appearance of Materials using Three-dimensional Textons", International Journal of Computer Vision, 2000. [pdf] [ps] [more info]
	A. A. Efros & T. K. Leung, "Texture Synthesis by Non-Parametric Sampling", International Conference on Computer Vision, Corfu, Greece, 1999. [pdf] [ps] [more info]
	T. K. Leung & J. Malik, "On Perpendicular Texture: Why do we see more flowers in the distance?", Conf. Computer Vision and Pattern Recognition, 1997 [pdf] [ps] [more info]
	T. K. Leung & J. Malik, "Detecting, Localizing and Grouping Repeated Scene Elements from an Image", European Conference on Computer Vision, Cambridge, England, 1996 [pdf] [ps] [more info]
	J. Malik & R. Rosenholtz, "Computing local surface orientation and shape from texture for curved surfaces." International Journal of Computer Vision 23(2):149-168. [pdf] [ps]