m4cheal / pcdcl-sfda

PRCV 2023: Pseudo-label Clustering-driven Dual-level Contrast Learning based Source-free Domain Adaptation for Fundus Image Segmentation

📄 Pseudo-label Clustering-driven Dual-level Contrast Learning based Source-free Domain Adaptation for Fundus Image Segmentation

Dependency Preparation

cd PCDCL_SFDA
# Python Preparation
conda create -n PCDCL_SFDA python=3.8.5
activate PCDCL_SFDA
# (torch 1.7.1+cu110) It is recommended to use the conda installation on the Pytorch website https://pytorch.org/
conda install pytorch==1.7.1 torchvision==0.8.2 torchaudio==0.7.2 cudatoolkit=11.0 -c pytorch
pip install -r requirements.txt

Model Training and Inference

• 1. Download the dataset and modify the relevant paths in the configuration file.
• 2. Source Model Train -- We use the code provided by ProSFDA to train the source model. If you want to use our trained source model, please contact me.
• 3. Generation phase: Generate target domain pseudo-labels

          python generate_pseudo.py

• 4. Adaptation stage: the source model adapts to the target domain

          python Train_target.py

Citation ✏️ 📄

If you find this repo useful for your research, please consider citing the paper as follows:

@inproceedings{zhou2023pseudo,
  title={Pseudo-Label Clustering-Driven Dual-Level Contrast Learning Based Source-Free Domain Adaptation for Fundus Image Segmentation},
  author={Zhou, Wei and Ji, Jianhang and Cui, Wei and Yi, Yugen},
  booktitle={Chinese Conference on Pattern Recognition and Computer Vision (PRCV)},
  pages={492--503},
  year={2023},
  organization={Springer}
}