This repository provides the code for paper: "FAS-UNet: A Novel FAS-driven Unet to Learn Variational Image Segmentation".
First, you can download the dataset at https://competitions.codalab.org/competitions/21145 (SegTHOR), or http://segchd.csail.mit.edu/index.html (HVSMR 2016), or https://chaos.grand-challenge.org/ (CHAOS CT)
Preprocess the dataset and save as ".npy" according to our paper. Split the preprocessed data into train data and validation data.
Place the ".npy" files in the "dataset_2D"(SegTHOR) or "01-chaos-overlap"(CHAOS CT) directory.
Run "main_mgunet_2d_374.py" or "main_fas_32_chaos_380_v2.py" to train and test 2D FAS-Unet or 3D FAS-Unet, respectively.