🥇 Winner solution on the MIPI 2022 Challenge on Under-display Camera Image Restoration

Our team (USTC_WXYZ) wins the MIPI 2022 Challenge on Under-display Camera Image Restoration!

Dependencies

• Python
• Pytorch (1.8+)
• Numpy

Testing on the Challendge Dataset

1. Prepare the testing data: put the testing data into './test_data' folder.

2. To test our model , run the command: "python testing.py"

3. Then, the processed results will be saved in './results' folder.

**Note that the pre-trained model has been saved in './ckpt' folder. And the pre-trained model is trained on the datasets privided by the challenge.

The testing results of the MIPI 2022 Challenge also could be downloaded from here.

Testing on the OTLED Dataset

We further training our scaled model on the OTLED dataset. The corresponding results could be found in here.

Thanks to the great efforts of the open-sourced projects MIMOUNet.

Citation

In case of use, please cite our publication:

@inproceedings{zhu2023enhanced,
  title={Enhanced Coarse-to-Fine Network for Image Restoration from Under-Display Cameras},
  author={Zhu, Yurui and Wang, Xi and Fu, Xueyang and Hu, Xiaowei},
  booktitle={Computer Vision--ECCV 2022 Workshops: Tel Aviv, Israel, October 23--27, 2022, Proceedings, Part V},
  pages={130--146},
  year={2023},
  organization={Springer}
}