Seeing Through the Rain: Resolving High-Frequency Conflicts in Deraining and Super-Resolution via Diffusion Guidance
1Beijing University of Posts and Telecommunications
2Nankai University
3Noah’s Ark Lab
⭐ If DHGM is helpful to your images or projects, please help star this repo. Thanks! 🤗
- Pytorch >= 1.8.1
- CUDA >= 11.1
- basicsr 1.4.2
# For install basicsr
pip install basicsr==1.4.2
python setup.py develop -i http://mirrors.aliyun.com/pypi/simple/
python -m pip install --upgrade pip
pip install numpy==1.24.4
pip install -v -e .
Download our processed training data from [Google Drive] to the input data folder. (Coming soon)
Download our processed testing data from [Google Drive] to the input data folder. (Coming soon)
You can download the qualitative results of our DHGM from [Google Drive].
# For Stage I
torchrun --nproc_per_node=$GPU_NUM$ basicsr/train.py -opt options/train_OursS1_x2_syn.yml --launcher pytorch
# For Stage II
torchrun --nproc_per_node=$GPU_NUM$ basicsr/train.py -opt options/train_OursS2_x2_syn.yml --launcher pytorch
Download the pretrained models from [Google Drive] to the experiments/Ours/models folder.
python basicsr/test.py -opt options/test_Ours_x2_syn.yml
If this work is helpful for your research, please consider citing the following BibTeX entry.
@inproceedings{li2026seeing,
title={Seeing Through the Rain: Resolving High-Frequency Conflicts in Deraining and Super-Resolution via Diffusion Guidance},
author={Li, Wenjie and Shi, Jinglei and Han, Jin and Guo, Heng and Ma, Zhanyu},
booktitle={AAAI},
year={2026}
}
The foundation for the training process is BasicSR, which profited from the outstanding contribution of XPixelGroup .
This repo is currently maintained by lewj2408@gmail.com and is for academic research use only.