This is the code repository for paper "A Novel Semi-supervised Training Framework Guided by Two Sources of Uncertainties for Medical Image Segmentation"

Training details about the model can be found in train_panc.py file in train function.

Folder data_lists specifies the train test split for pancreas dataset and left atrium dataset.

Folder preprocess contains code used to preprocess pancreas dataset. If you are using raw pancreas dataset, you may need to preprocess data first by running pancreas_preprocess.py.

Folder trained_models contains the model whose results are presented in our paper.

Files with suffix "panc" means it uses pancreas dataset while suffix "LA" means left atrium dataset

If you'd like to train the model from scratch, you can run either train_panc.py or train_LA.py. You may need uncomment the line invoking pretrain function to get a pretrained model first and prepare corresponding datasets and modify the dataset path in the code.

Trained models on pancreas and left atrium dataset is available on google drive