Imaging gravity-induced lung water redistribution with automated inline processing at 0.55T MRI

This code is for the manuscript (in sub): Seemann, F., et al. Imaging gravity-induced lung water redistribution with automated inline processing at 0.55T MRI

This repository contains

• Gadgetron and Matlab code for inline reconstruction and analysis
• Python processing for lung segmentation for analysis, including a pkl file from human and pig data.
• Segment plugin for offline analysis

NHLBI TOOLBOX GADGETRON Installation

Repository with gadgetron image reconstruction code: https://github.com/NHLBI-MR/selfgated_noncartesian_reconstruction.git

% GADGETRON MATLAB

• Install application through matlab
• https://www.mathworks.com/matlabcentral/fileexchange/72715-gadgetron
• Make directory "+nhlbi" inside the +gadgetron class
• Copy the lung_water_pipeline.m in to the +nhlbi folder.

Reconstruction & Inline Lung Water Analysis

• gadgetron configuration file

Offline analysis using Segment

Installation

• https://github.com/Cardiac-MR-Group-Lund/segment-open
• Medviso

Requires python packages

• numpy
• matplotlib
• torch
• torchvision
• scipy

pip install numpy matplotlib torch torchvision scipy the trained neural network is named best_lung_seg_model.pkl and is saved in the python folder