https://github.com/masilab/beyond_fa_microstruct_baseline
Microstructural baseline for BeyondFA challenge
https://github.com/masilab/beyond_fa_microstruct_baseline
Last synced: 11 days ago
JSON representation
Microstructural baseline for BeyondFA challenge
- Host: GitHub
- URL: https://github.com/masilab/beyond_fa_microstruct_baseline
- Owner: MASILab
- License: mit
- Created: 2025-01-31T23:12:04.000Z (12 months ago)
- Default Branch: main
- Last Pushed: 2025-03-20T20:41:14.000Z (10 months ago)
- Last Synced: 2025-03-20T22:02:51.941Z (10 months ago)
- Language: Shell
- Homepage:
- Size: 455 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE.md
Awesome Lists containing this project
README
# Beyond FA Baseline
## Challenge overview
We provide the overall structure of our challenge below. Participants are only responisble for the green "Participant Docker" oval -- everything else is provided on the [Beyond FA grand-challenge site](https://bfa.grand-challenge.org/).
![alt text][ChallengeArchitecture]
[ChallengeArchitecture]: https://github.com/MASILab/beyondFA_mlp/blob/main/fig2.png "Challenge Architecture"
Here is an example repository for the Beyond FA challenge. This model calculates fractional anisotropy (FA) from the diffusion MRI, then finds the average value of FA in regions defined by [TractSeg](https://github.com/MIC-DKFZ/TractSeg).
## Building the Docker
See `Dockerfile` for an example of setting up the Docker container. The Docker container is based the mrtrix Docker, which includes several common neuroimaging packages (e.g., MRtrix3, ANTs, FSL, Freesurfer). It also installs the following:
- [`uv`](https://github.com/astral-sh/uv) for installing and running Python in a virtual environment
- [`TractSeg`](https://github.com/MIC-DKFZ/TractSeg)
- [`scilpy`](https://github.com/scilus/scilpy)
To build this Docker container, clone the repository and run the following command in the root directory:
```bash
DOCKER_BUILDKIT=1 sudo docker build -t beyondfa_baseline:v1.1.8 .
```
The Docker runs the code from `scripts/entrypoint.sh`.
## Running the Docker
Your Docker container should be able to read input data from `/input` and write output data to `/output`. Intermediate data should be written to `/tmp`. The input data will be a `.mha` file containing the diffusion MRI data with gradient table information contained in a `.json` file. The input file will be in `/input/images/dwi-4d-brain-mri/`, with gradient table information at `/input/dwi-4d-acquisition-metadata.json`. Your Docker should write a JSON list to the output directory with the name `/output/features-128.json`. **Your JSON list must contain 128 values. You may zero-pad the list if you wish to provide fewer than 128 values.**
See `scripts/convert_mha_to_nifti.py` and `scripts/convert_json_to_bvalbvec.py` for scripts to convert the `.mha` to `.nii.gz` and the `.json` to `.bval` and `.bvec` files.
To run this Docker:
```bash
input_dir=".../input_data"
output_dir=".../output_data"
mkdir -p $output_dir
DOCKER_NOOP_VOLUME="beyondfa_baseline-volume"
sudo docker volume create "$DOCKER_NOOP_VOLUME" > /dev/null
sudo docker run \
-it \
--platform linux/amd64 \
--network none \
--gpus all \
--rm \
--volume $input_dir:/input:ro \
--volume $output_dir:/output \
--volume "$DOCKER_NOOP_VOLUME":/tmp \
beyondfa_baseline:v1.1.8
sudo docker volume rm "$DOCKER_NOOP_VOLUME" > /dev/null
```