Prepare data

This page is a bookmark. Load images and masks into a Cohort, then run a map recipe. Which gallery page to copy depends on the files you already have (1. Data In):

• HABIT directory images/<subject>/<series>/ and masks/<subject>/<roi>/ — Load from directory

• Loose NIfTI / NRRD / MetaImage, one subject or many, series name different from the ROI name, or grids that differ — Load from NIfTI files

• SimpleITK images already in memory — Load from SimpleITK

• NumPy arrays or tensors — Load from NumPy arrays

Recipes take that cohort. cohort[:1] stays a cohort. A Python list of subjects does not. DICOM still goes through Preprocessing before any of the routes above.

Then run a map recipe: Two-step habitat analysis, One-step habitat analysis, or Apply a saved model.