This repository contains the implementation of the paper Constructing and Exploring Intermediate Domains in Mixed Domain Semi-supervised Medical Image Segmentation

In Conference on Computer Vision and Pattern Recognition (CVPR), 2024

The dataset needs to be divided into two folders for training and testing. The training and testing data should be in the format of the "data" folder.

code/train.py is the implementation of our method on the Prostate and Fundus dataset.

code/train_MNMS.py is the implementation of our method on the M&Ms dataset.

Modify the paths in lines 631 to 636 of the code.

if args.dataset == 'fundus':
    train_data_path='../../data/Fundus' # the folder of fundus dataset
elif args.dataset == 'prostate':
    train_data_path="../../data/ProstateSlice" # the folder of prostate dataset
elif args.dataset == 'MNMS':
    train_data_path="../../data/MNMS/mnms" # the folder of mnms dataset

then simply run:

python train.py --dataset ... --lb_domain ... --lb_num ... --save_name ... --gpu 0

To run the evaluation code, please update the path of the dataset in test.py:

Modify the paths in lines 249 to 254 of the code.

then simply run:

python test.py --dataset ... --save_name ... --gpu 0

Prostate with the extraction code: 4no2

Fundus

M&Ms with the extraction code: cdbs

The Prostate and M&Ms datasets have undergone preprocessing in our work, with the original data sourced from Prostate and M&Ms

This project is based on the code from the SSL4MIS project.

Thanks a lot for their great works.