/predictTCL

Implementation for Joint Prediction of Meningioma Grade and Brain Invasion via Task-Aware Contrastive Learning

Primary LanguagePython

predictTCL

Implementation for "Joint Prediction of Meningioma Grade and Brain Invasion via Task-Aware Contrastive Learning"[https://arxiv.org/pdf/2209.01517] in pytorch. MICCAI 2022 accept.

Requirements

  • Python 3
  • PyTorch
  • SimpleITK
  • scikit-learn
  • scipy

File structure

-sup

-----contrastive_loss.py

-----data_provider.py

-----make_randomfolders.py

-----online_aug.py

-----predictTCL.py

-----resnet3d.py

-----scheduler.py

-main.py

Running the Code

-Firstly, run make_randomfolders.py to make 3 random folder data.

-Then， run python main.py to train and test. Change the parameter ‘run_type’ in main.py to ‘train’ for train and ‘test’ for test.