ImageTypes

Univariate Images

Univariate images are m-by-n-by-1 (where m and n are pixels) array where each value in the array represents a pixel intensity value.

Multivariate Image 3 Variables

Most digital photos are "Truecolor" or "RGB" (Red/Green/Blue) and come in the form or m-by-n-by-3 where there are 3 "slabs" of pixel intensities.

Multivariate Image 3+ Variables

Images with more than 3 variables are often referred to as hyperspectral images. These arrays will be m-by-n-by-k where k might be select wavelengths or an entire spectrum for each m/n pixel.

Volumetric And Other Image Sizes

Note that the descriptions above describe the image as having two spatial dimensions and one dimension of measured variables. It is possible to have volumetric 3D images in which there are 3 spatial dimensions in conjunction with a dimension of measured variables, or even 4D images (e.g. 3 spatial dimensions and a time dimension) in conjunction with a dimension of measured variables.

It is also possible to have multiple dimensions of measured variables in conjunction with any of these. For example, an experiment like Excitation Emission Fluorescence spectroscopy measures intensity as a function of 2 variables (Excitation wavelength and Emission wavelength). If this is measured at every position in a 3D volume, you could have 3 spatial dimensions and 2 variable dimensions for a total of a 5-dimensional DataSet. The instructions for Constructing Image DataSets applies to these higher-dimension data types as well.