FDAS: Foundation Model Distillation and Anatomic Structure-aware Multi-task Learning for Self-Supervised Medical Image Segmentation

---

This repository provides the code for our MICCAI 2025 paper "FDAS: Foundation Model Distillation and Anatomic Structure-aware Multi-task Learning for Self-Supervised Medical Image Segmentation"

✨Overall Framework

🔨Usage

Download dataset

1. Download the M&MS Dataset, and organize the dataset directory structure as follows.

your/data_root/
       train/
            img/
                full/      # 100% of the training images
                    A0S9V9_0.nii.gz
                    ...
                subset20/  # 20% of the training images
                subset10/  # 10% of the training images
                ...
            lab/
                full/
                    A0S9V9_0_gt.nii.gz
                    ...
                subset20/
                subset10/
                ...
       valid/
            img/
            lab/
       test/
           img/
           lab/

Download SAM model

2. Download the SAM model and move the model to the "your_root/pretrained_model" directory in your project.

Pre-processing data

3. Generate masks using the Segment Anything Model (SAM)

generate_masks.py

4. Pre-process the masks generated from SAM

pre_process.py

Pre-training

5. Pre-train the network using pre_train.py

pre_train.py

Fine-tuning

6. Fine-tune the network using fine_tune.py

fine_tune.py

📝 Citation

If you find this project useful for your research, please consider citing:

@inproceedings{qi2025fdas,
  title={FDAS: Foundation Model Distillation and Anatomic Structure-Aware Multi-task Learning for Self-Supervised Medical Image Segmentation},
  author={Qi, Xiaoran and Zhang, Guoning and Wu, Jianghao and Zhang, Shaoting and Hou, Xiaorong and Wang, Guotai},
  booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
  pages={192--202},
  year={2025},
  organization={Springer}
}

🤝 Acknowledgement

• Parts of our codebase are adapted from PyMIC. We also thank the authors of the Segment Anything project for open-sourcing their code.