Code for implementing and testing the Generalized Surface Loss (GSL).
MICCAI Liver and Tumor Segmentation Challenge 2017 dataset (LiTS) - https://competitions.codalab.org/competitions/17094#learn_the_details-overview
MICCAI Brain Tumor Segmentation Challenge 2020 dataset (BraTS) - https://www.med.upenn.edu/cbica/brats2020/registration.html
This code is based on the Medical Imaging Segmentation Toolkit (MIST) framework. Please see its documentation for setup details.
To run our experiments, use the following commands:
python run_experiments_lits.py
and
python run_experiment_brats.py