AdaMix

This repo is the PyTorch implementation of the paper:

"Adaptive Mix for Semi-Supervised Medical Image Segmentation"

Usage

0. Requirements

The code is developed using Python 3.8 with PyTorch 1.11.0, and CUDA 11.3. All experiments in our paper were conducted on a single NVIDIA A40 GPU with 48GB memory.

Install the main packages:

pytorch == 1.11.0
torchvision == 0.12.0
cudatoolkit == 11.3.1

1. Data Preparation

1.1. Download data

The original data can be downloaded in following links:

• ACDC Dataset - Link (Original) - Link (Processed)
• LA Dataset - Link (Original) Link (Processed)
• ISIC Dataset - Link

PS: Please cite the papers of original datasets when using the data in your publications.

1.2. Split Dataset

Following the list files (within the "data" folders) to split the datasets.

2. Training

python train_adamix_[st/mt/ct].py

3. Evaluation

python eval.py

Citation

If you find this project useful, please consider citing:

@article{shen2026adaptive,
  title={Adaptive Mix for Semi-Supervised Medical Image Segmentation},
  author={Shen, Zhiqiang and Cao, Peng and Su, Junming and Yang, Jinzhu and Zaiane, Osmar R},
  journal = {Medical Image Analysis},
  volume = {108},
  pages = {103857},
  year = {2026},
  doi = {https://doi.org/10.1016/j.media.2025.103857}
}

Contact

If you have any questions or suggestions, please feel free to contact me (xxszqyy@gmail.com).