MRI_reface: face-off of MRI research sequences

In neuroimage research, images may potentially be re-identified using face recognition.

“de-facing” images is becoming a part of HPC workflows. On ALICE mri_reface

( https://doi.org/10.1016/j.neuroimage.2023.120199 ) is one such tool.

Highlights
•3D T1-w, 3D T2-w, and 3D T2-FLAIR were all highly re-identifiable (96–98%).
•2D T2-FLAIR and 3D multi-echo GRE were each moderately re-identifiable (44–45%).
•Diffusion, functional and ASL were each minimally re-identifiable (0–8%).
•mri_reface reduced all re-identification rates to ≤8%.
•mri_reface effects on brain measurements were all below or similar to scan-rescan.

A module is available to facilitate running the (apptainer) container and wrapper script.

Interactively (on a compute node)

Full options with “run_mri_reface -h”

TODO slurm script