DL-Generic : Internal project at Grenoble Institut of Neuroscience

Installation:

First dispatch your data in a container folder "$data" in 3 folder "$train" "$val" "$test", using the BIDS nomenclature $data/$train/$subject_id/${subject_id}_${sequences_name}.nii.gz
Then generate patches to feed our neural network using the patches_generation.py script. --help will list required or optional additional parameters
Then you can generate and train you model using the "use_model.py" script. --help will list required or optional additional parameters
In case of generation / segmentation usage you might need to reconstruct infered output using the reconstruction.py script

Name		Name	Last commit message	Last commit date
Latest commit History 5 Commits
.gitignore		.gitignore
CNN.py		CNN.py
ClassificationTransformer.py		ClassificationTransformer.py
DualGAN.py		DualGAN.py
GAN.py		GAN.py
ImUnity.py		ImUnity.py
Markovian_CNN.py		Markovian_CNN.py
Models.py		Models.py
PatchGAN.py		PatchGAN.py
README.md		README.md
Resnet.py		Resnet.py
SpectralNormalizationKeras.py		SpectralNormalizationKeras.py
Transformer.py		Transformer.py
Unet.py		Unet.py
VAE.py		VAE.py
VAEGAN.py		VAEGAN.py
augmentations.py		augmentations.py
cross_val.sh		cross_val.sh
get_cross_val_groups.py		get_cross_val_groups.py
get_test_accuracy.py		get_test_accuracy.py
layers.py		layers.py
ops.py		ops.py
patches_generation.py		patches_generation.py
reconstruction.py		reconstruction.py
requirements.txt		requirements.txt
sn.py		sn.py
use_model.py		use_model.py
utils.py		utils.py