Skip to content

Latest commit

 

History

History
80 lines (65 loc) · 2.69 KB

README.md

File metadata and controls

80 lines (65 loc) · 2.69 KB
Raw

Self-supervised Non-uniform Kernel Estimation with Flow-based Motion Prior for Blind Image Deblurring (CVPR 2023)

Zhenxuan Fang, Fangfang Wu, Weisheng Dong, Xin Li, Jinjian Wu and Guangming Shi

[Paper] [Website Page]

Architecture

Fig. 1 The overall framework of the proposed UFPNet for blind SR.

Usage

This implementation based on BasicSR and NAFNet

Download the repository

  1. Requirements Python 3.7 and PyTorch 1.8.0.
  2. Download this repository via git
git clone https://github.com/Fangzhenxuan/UFPDeblur

or download the zip file manually.

Quick Start

Download the pretrained checkpoints (Google Drive), the directory structure will be arranged as:

experiments
    |- pretrained_models
        |- train_on_GoPro
        |- train_on_RealBlurJ
        |- train_on_RealBlurR

Put the test datasets in dir ./datasets/

datasets
    |- GoPro
        |- test
            |- target
            |- input
    |- ...
  • Test on GoPro testset, run 
    python ./basicsr/test.py -opt options/test/GoPro/UFPNet-GoPro.yml
  • Test on RealBlur-J testset
    • To use the model trained on GoPro, run 
      python ./basicsr/test.py -opt options/test/RealBlur-J/UFPNet-RealBlurJ-Train-on-GoPro.yml
    • To use the model trained on RealBlur-J, run 
      python ./basicsr/test.py -opt options/test/RealBlur-J/UFPNet-RealBlurJ.yml

Citations

If UFPNet helps your research or work, please consider citing UFPNet.

@inproceedings{fang2023self,
  title={Self-supervised Non-uniform Kernel Estimation with Flow-based Motion Prior for Blind Image Deblurring},
  author={Fang, Zhenxuan and Wu, Fangfang and Dong, Weisheng and Li, Xin and Wu, Jinjian and Shi, Guangming},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={18105--18114},
  year={2023}
}

Acknowledgements

The codes are built on NAFNet [1]. We thank the authors for sharing their codes.

References

[1] Liangyu Chen, et al. "Simple Baselines for Image Restoration." In European Conference on Computer Vision 2022.