This repository is Pytorch code for our proposed uncertainty-driven loss (UDL).
The code is built on RCAN and tested on Ubuntu 16.04 environment (Python 3.5/3.6/3.7, PyTorch 1.4.0) with 2080Ti/1080Ti GPUs.
If you find our work useful in your research or publications, please consider citing:
@inproceedings{ning2021uncertainty,
title={ Uncertainty-Driven Loss for Single Image Super-Resolution },
author={ Ning Qian and Dong, WeiSheng and Li, Xin and Wu, Jinjian and Shi, Guangming },
booktitle={Advances in Neural Information Processing Systems},
year={2021}
}
- Python 3
- skimage
- imageio
- Pytorch (Pytorch version 1.0.1 is recommended)
- tqdm
- cv2 (pip install opencv-python)
cd code
sh train.sh
-
If you have cloned this repository, the pre-trained models can be found in experiment fold and test dataset Set5 can be found in data fold.
-
Then, run command:
cd code sh test.sh
-
Finally, PSNR values are shown on your screen, you can find the reconstruction images in
../experiment/xx/results/
- This code is built on RCAN (PyTorch). We thank the authors for sharing their codes.