radiology-preprocessing

Installation
SKILL.md

Radiology Preprocessing

Overview

Deterministic commands for preprocessing structural brain MRI (T1w, T2w, FLAIR) and related volumetric radiology data. Covers the canonical pipeline used by most downstream analyses (segmentation, registration studies, deep-learning training): reorient → bias correction → skull strip → registration → intensity normalization.

Tools used:

  • HD-BET — CNN-based brain extraction (state-of-the-art on pathological brains).
  • FSLbet, flirt, fnirt, fslmaths, fslreorient2std.
  • ANTsN4BiasFieldCorrection, antsRegistrationSyNQuick.sh, antsApplyTransforms.
  • FreeSurfermri_convert for resampling/format conversion.

All commands operate on NIfTI (.nii / .nii.gz). Run in a POSIX shell with the respective tool on PATH.

Usage

Standard T1w preprocessing pipeline

Run these steps in order on a single subject volume:

Installs
6
GitHub Stars
11
First Seen
Jul 9, 2026
radiology-preprocessing — awslabs/hcls-agent-skills