🗓 2025.03.19: Upload the code for Glaucoma&Diabetic classification.
🗓 2025.03.06: Repository created.
- Extend the experiment on segmentation task.
- Code will be released soon. ⏳
git clone git@github.com:anonymous/DiffCTA.git
cd DiffCTAconda create -n DiffCTA python=3.8 -y
conda activate DiffCTA
pip3 install -r requirements.txt - Download the dataset and Checkpoint using the following command:
wget https://oneflow-static.oss-cn-beijing.aliyuncs.com/data_lx/Fundus.zipbash optic_adapt.sh
- Run the following command to perform a quick inference:
```bash
bash TTA.sh
| Method | Domain A | Domain B | Domain C | Domain D | Domain E | AVG |
|---|---|---|---|---|---|---|
| Source Only | 68.37 | 50.68 | 65.74 | 34.98 | 43.43 | 52.64 |
| TENT | 65.84 | 58.84 | 60.36 | 28.42 | 42.76 | 51.24 |
| CoTTA | 64.01 | 58.51 | 61.25 | 24.02 | 33.59 | 48.28 |
| EATA | 66.46 | 58.50 | 63.34 | 33.41 | 40.42 | 52.43 |
| SAR | 66.57 | 58.81 | 63.21 | 32.87 | 33.52 | 51.00 |
| DDA | 69.71 | 56.06 | 67.20 | 34.22 | 39.73 | 53.38 |
| DiffCTA | 70.47 | 59.34 | 68.46 | 35.59 | 45.93 | 55.96 |
The code and models are licensed under MIT License.