MeasTex Image Texture Database and Test Suite
Version 1.1

Updated 27 May, 1997


MeasTex is an image database and quantitative measurement framework for image texture analysis algorithms.

MeasTex should be of general interest to the image processing community, and of specific interest to the image texture analysis community.

Briefly

MeasTex provides:

Contents

We have provided an introduction to MeasTex for readers with the following interests:

Version 1.1 of the MeasTex framework is now available. See the Maintenance log for bug fixes and other changes.

What is MeasTex?

MeasTex is about the MEASurement of TEXture classification algorithms.

Visual texture is a property of a region in an image. Texture is independent of the colour and brightness of the region. The grain of timber is a good example of image texture. Different species of timber have visually distinct grain patterns. These patterns allow these species to be identified, even if the natural colour of the wood has been altered by staining, aging or treatment with chemicals.

Visual texture, or image texture, has many important scientific applications. At a large scale, image texture has been used in the analysis of satellite images. At a much smaller scale, the texture of cell nuclei are among the best indicators of precancerous cells in medical screening programs.

The computerised recognition of image textures is a long standing field of research. As would be expected in a long standing field, many different paradigms for the analysis of texture have been proposed. However, only recently have large databases of textured images been available on the InterNet, and comparatively few quantitative comparisons of algorithms have been published.

MeasTex is a framework for quantitative comparisons of texture classification algorithms. This site contains a database of images captured specificly for image texture analysis. This site also contains the formats, software and test suites necessary to measure an algorithm in the MeasTex framework, and implementations of and results for some well known texture classification algorithms.




Return to TextureSynthesis Homepage


Guy Smith guy@it.uq.edu.au
Ian Burns burns@it.uq.edu.au

Last Modified: Tue May 27 17:34:33 EST 1997