Motivations

A generic algorithm written to manipulate a medical image saved in nifti format usually involves these steps:

1. loading the image with nibabel
2. extract the header or the data
3. apply the intended algorithm (usually numpy or scipy based).
4. create a new instance of the modified nifti image
5. save the modified image

NiLabels is aimed at reducing these steps in a single command, allowing to have as input and output directly the path to the input and output image.

• The underpinning algorithm is written (and sometimes tested) for numpy arrays.
• The facade accesses the underpinning algorithm providing directly the path to the input image.

Nilabels had begun within the GIFT-surg research project, by the collaboration with Tom Vercauteren, Johannes van der Merwe, Jan Deprest and Sebastiano Ferraris. It originated as a component of the data analysis pipelines developed during a study for the analysis of the encephalopathy of prematurity in the newborn rabbit brain.