Sign inSign up

mahdisabetkish/brainssl

By mahdisabetkish

•Updated 9 days ago

Brain MRI tumour segmentation and fMRI autism classification: training images (CPU and CUDA 12.6).

Image
0

673

mahdisabetkish/brainssl repository overview

⁠brainssl

Label-efficient learning on brain MRI and resting-state fMRI: tumour segmentation with a bilateral 3D U-Net (MSD Task01) and autism classification from connectomes with a noise-aware linear model (ABIDE I).

Code, documentation and the interactive dashboard: https://github.com/mahdisabetkish/brainssl⁠ Dashboard: https://mahdisabetkish.github.io/brainssl/⁠

⁠Tags

TagContentsSize
cpu, latestPython 3.11, PyTorch (CPU), the brainssl package. Enough for the connectome study, the tests and the synthetic pipeline testabout 0.5 GB
gpuSame package with PyTorch for CUDA 12.6, for segmentation trainingabout 3.3 GB

⁠Use

docker pull mahdisabetkish/brainssl:cpu
docker run --rm mahdisabetkish/brainssl:cpu brainssl --help

# with a GPU and your data folder mounted
docker run --rm --gpus all -v "$PWD:/work" -w /work mahdisabetkish/brainssl:gpu brainssl prepare msd

The images hold code only. Data are downloaded with brainssl download, and the licences of the datasets (CC BY-SA 4.0 for MSD, CC BY-NC-SA for ABIDE) apply to them. Research use only, not a medical device.

Tag summary

Content type

Image

Digest

sha256:a79ba3603…

Size

447.5 MB

Last updated

9 days ago

docker pull mahdisabetkish/brainssl