XPHT allows the computation of the extended persistent homology transform of binary images detailed in Turner, Robins, & Morgan (2022). This provides a simple method of computing the Wasserstein $p$ distance between image for statistical shape analysis.

This package relies on the Imager package. This allows us to read in images in order to extract the boundary curves. There is detailed information about installing this package, so we direct the user to view that if any difficulties are encountered.

If you wish to run the development version, install the devtools package (if you haven’t already) and then run:

devtools::install_github("james-e-morgan/xpht")

XPHT requires the input be binary images. For this it is required that the pixel values of the input images are either 0 or 1. If any other values are present an error will be thrown.