A tool for the automatic segmentation of fetal brain 3D T2-weighted MRI.
To run the automatic segmentation algorithms a NVIDIA GPU with at least 8GB of memory is required.
The code has been tested with the configuration:
- 12 Intel(R) Core(TM) i7-8750H CPU @ 2.20GHz
- 1 NVIDIA GPU GeForce GTX 1070 with 8GB of memory
The code is supported on every operating system (OS) using docker. It has been tested for
- Linux Ubuntu 18.04.6 LTS
The installation is performed using docker:
First, install docker (see https://docs.docker.com/get-docker/).
Download the latest docker image using
docker pull lucasfidon/fetal_brain_segmentation:0.1
You can compute the automatic segmentations for a fetal brain 3D T2w MRI using the docker image downloaded above.
To learn more about the usage of the docker image, please see
docker run --rm lucasfidon/fetal_brain_segmentation:0.1 -h
We refer to the demo below for a detailed example.
Fetal brain 3D MRI from a subset of the training dataset can be downloaded at https://zenodo.org/record/6405632#.YkbWPCTMI5k
Unzip the archive FeTA2021_Release1and2Corrected_v2.zip
Go to the folder containing the 3D MRI examples
cd FeTA2021_Release1and2Corrected_v2
We will segment the 3D T2w MRI contained in the folder sub-041
.
The folder contains:
srr.nii.gz
: the 3D T2w MRI to segment. This is the main input of the segmentation algorithm.mask.nii.gz
: the brain mask forsrr.nii.gz
. This is the second input of the segmentation algorithm.parcellation.nii.gz
: the manual segmentation forsrr.nii.gz
. After computing the automatic segmentation you can compare it to this segmentation.
Create a folder for the results of the automatic segmentation algorithm for the case sub-041
mkdir results-sub-041
Run the automatic segmentation for the case sub-041/srr.nii.gz
using
docker run -v <absolute-path-to-sub-041>:/input -v <absolute-path-to-results-sub-041>:/output --gpus 0 --rm lucasfidon/fetal_brain_segmentation:0.1 --img /input/srr.nii.gz --mask /input/mask.nii.gz --output_folder /output
This will take between 1 minute and 3 minutes.
The automatic segmentation will be in saved in results-sub-041/srr_segmentation_AI.nii.gz
.
If you find this repository useful for your research, please consider giving us a star ⭐ and cite
- L. Fidon, M. Aertsen, F. Kofler, A. Bink, A. L. David, T. Deprest, D. Emam, F. Guffens, A. Jakab, G. Kasprian, P. Kienast, A. Melbourne, B. Menze, N. Mufti, I. Pogledic, D. Prayer, M. Stuempflen, E. Van Elslander, S. Ourselin, J. Deprest, T. Vercauteren. A Dempster-Shafer approach to trustworthy AI with application to fetal brain MRI segmentation
@article{fidon2022dempster,
title={A Dempster-Shafer approach to trustworthy AI with application to fetal brain MRI segmentation},
author={Fidon, Lucas and Aertsen, Michael and Kofler, Florian and Bink, Andrea and David, Anna L and Deprest, Thomas and Emam, Doaa and Guffens, Fr and Jakab, Andr{\'a}s and Kasprian, Gregor and others},
journal={arXiv preprint arXiv:2204.02779},
year={2022}
}