What is PAT-GEOM?
PAT-GEOM is a free-to-use program written by Ian Z.W. Chan (and published in Methods in Ecology and Evolution) that measures 7 properties of pattern: shape, shape directionality, size, contrast, distribution of markings, distribution directionality and randomness.
Colour patterns can be decomposed into (1) colour, (2) polarity and (3) the underlying pattern, i.e. pattern per se. Whilst there are many techniques to quantify colour (e.g. using RGB values or reflectance spectra), the same is not true for the underlying pattern. PAT-GEOM is our attempt to address this deficiency.
Although designed for animal colour patterns (e.g. zebra stripes), it could potentially also be used in botany (to measure leaf shape), landscape ecology (to describe landscape patches), cell biology (to measure chromosome shape and alignment), molecular biology (to describe the distribution of RNA or protein species) and in many other fields, so long as a spatial pattern is to be measured.
The measurements produced by PAT-GEOM are fully quantitative and could be used for statistical analyses such as t-tests, generalised linear models, PCA and heat maps.
More information on how the properties are measured, detailed instructions on each step (from installation to analysis) and Worked Examples with sample output data can be found in the User Guide and in the published article.
Finally, the programs are designed to work as macros in ImageJ (and so require ImageJ to be installed ;see the PAT-GEOM User Guide for instructions). They work with any image format that ImageJ can open, but were designed specifically with multispectral (mspec) images in mind. These mspec images are part of the Multispectral Image Calibration and Analysis (MICA) toolbox created by Jolyon Troscianko & Martin Stevens (2015). Their technique is slightly involved but well worth the time and effort to learn because it is a rigorous way of analysing colour patterns across the human visible and UV range.