Skip to content
Branch: master
Find file History
Latest commit c4dc8bb Jan 12, 2019
Permalink
Type Name Latest commit message Commit time
..
Failed to load latest commit information.
RDN_CAR_Y RDN for IR Jan 11, 2019
RDN_DN_Gray RDN for IR Jan 11, 2019
RDN_DN_RGB RDN for IR Jan 11, 2019
RDN_SR_RGB RDN for IR Jan 11, 2019
Readme.md Update Readme.md Jan 12, 2019

Readme.md

Residual Dense Network for Image Restoration

This repository is for RDN introduced in the following papers

Yulun Zhang, Yapeng Tian, Yu Kong, Bineng Zhong, and Yun Fu, "Residual Dense Network for Image Super-Resolution", CVPR 2018 (spotlight), [arXiv]

Yulun Zhang, Yapeng Tian, Yu Kong, Bineng Zhong, and Yun Fu, "Residual Dense Network for Image Restoration", arXiv 2018, [arXiv]

The code is built on EDSR (Torch) and tested on Ubuntu 14.04 environment (Torch7, CUDA8.0, cuDNN5.1) with Titan X/1080Ti/Xp GPUs.

Other implementations: PyTorch_version has been implemented by Nguyễn Trần Toàn (trantoan060689@gmail.com) and merged into EDSR_PyTorch. TensorFlow_version by hengchuan.

To be updated!!!

Contents

  1. Test
  2. Results
  3. Citation
  4. Acknowledgements

Test

Quick start for SR (BI degradation model, training data: DIV2K+Flickr2K, input size: 48x48)

  1. Download models for our paper and place them in '/RDN_IR/RDN_TestCode/RDN_SR_RGB/model'.

    All the models can be downloaded from Baidu.

  2. Run 'TestRDN.lua'

    You can use the following scripts to produce results for our paper.

    # No self-ensemble: RDN
    # BI degradation model, X2, X3, X4, X8
    th TestRDN.lua -model RDN_DF2K_BIX2 -degradation BI -scale 2 -selfEnsemble false -dataset Set5
    th TestRDN.lua -model RDN_DF2K_BIX3 -degradation BI -scale 3 -selfEnsemble false -dataset Set5
    th TestRDN.lua -model RDN_DF2K_BIX4 -degradation BI -scale 4 -selfEnsemble false -dataset Set5
    th TestRDN.lua -model RDN_DF2K_BIX8 -degradation BI -scale 8 -selfEnsemble false -dataset Set5
    
    # With self-ensemble: RDN+
    # BI degradation model, X2, X3, X4, X8
    th TestRDN.lua -model RDN_DF2K_BIX2 -degradation BI -scale 2 -selfEnsemble true -dataset Set5
    th TestRDN.lua -model RDN_DF2K_BIX3 -degradation BI -scale 3 -selfEnsemble true -dataset Set5
    th TestRDN.lua -model RDN_DF2K_BIX4 -degradation BI -scale 4 -selfEnsemble true -dataset Set5
    th TestRDN.lua -model RDN_DF2K_BIX8 -degradation BI -scale 8 -selfEnsemble true -dataset Set5

Quick start for DN_Gray

  1. Download models for our paper and place them in '/RDN_IR/RDN_TestCode/RDN_DN_Gray/model'.

    All the models can be downloaded from Baidu.

  2. Run 'TestRDN_DN_Gray.lua'

    You can use the following scripts to produce results for our paper.

    # No self-ensemble: RDN
    th TestRDN_DN_Gray.lua -model RDN_DN_Gray_N10 -noise 10 -selfEnsemble false -dataset Kodak24 
    th TestRDN_DN_Gray.lua -model RDN_DN_Gray_N30 -noise 30 -selfEnsemble false -dataset Kodak24 
    th TestRDN_DN_Gray.lua -model RDN_DN_Gray_N50 -noise 50 -selfEnsemble false -dataset Kodak24 
    th TestRDN_DN_Gray.lua -model RDN_DN_Gray_N70 -noise 70 -selfEnsemble false -dataset Kodak24 
    
    # With self-ensemble: RDN+
    th TestRDN_DN_Gray.lua -model RDN_DN_Gray_N10 -noise 10 -selfEnsemble true -dataset Kodak24 
    th TestRDN_DN_Gray.lua -model RDN_DN_Gray_N30 -noise 30 -selfEnsemble true -dataset Kodak24 
    th TestRDN_DN_Gray.lua -model RDN_DN_Gray_N50 -noise 50 -selfEnsemble true -dataset Kodak24 
    th TestRDN_DN_Gray.lua -model RDN_DN_Gray_N70 -noise 70 -selfEnsemble true -dataset Kodak24 

Quick start for DN_RGB

  1. Download models for our paper and place them in '/RDN_IR/RDN_TestCode/RDN_DN_RGB/model'.

    All the models can be downloaded from Baidu.

  2. Run 'TestRDN_DN_RGB.lua'

    You can use the following scripts to produce results for our paper.

    # No self-ensemble: RDN
    th TestRDN_DN_RGB.lua -model RDN_DN_RGB_N10 -noise 10 -selfEnsemble false -dataset Kodak24 
    th TestRDN_DN_RGB.lua -model RDN_DN_RGB_N30 -noise 30 -selfEnsemble false -dataset Kodak24 
    th TestRDN_DN_RGB.lua -model RDN_DN_RGB_N50 -noise 50 -selfEnsemble false -dataset Kodak24 
    th TestRDN_DN_RGB.lua -model RDN_DN_RGB_N70 -noise 70 -selfEnsemble false -dataset Kodak24 
    
    # With self-ensemble: RDN+
    th TestRDN_DN_RGB.lua -model RDN_DN_RGB_N10 -noise 10 -selfEnsemble true -dataset Kodak24 
    th TestRDN_DN_RGB.lua -model RDN_DN_RGB_N30 -noise 30 -selfEnsemble true -dataset Kodak24 
    th TestRDN_DN_RGB.lua -model RDN_DN_RGB_N50 -noise 50 -selfEnsemble true -dataset Kodak24 
    th TestRDN_DN_RGB.lua -model RDN_DN_RGB_N70 -noise 70 -selfEnsemble true -dataset Kodak24 

Quick start for CAR_Y

  1. Download models for our paper and place them in '/RDN_IR/RDN_TestCode/RDN_CAR_Y/model'.

    All the models can be downloaded from Baidu.

  2. Run 'TestRDN_CAR_Y.lua'

    You can use the following scripts to produce results for our paper.

    # No self-ensemble: RDN
    th TestRDN_CAR_Y.lua -model RDN_CAR_Y_Q10 -noise 10 -selfEnsemble false -dataset Classic5
    th TestRDN_CAR_Y.lua -model RDN_CAR_Y_Q20 -noise 20 -selfEnsemble false -dataset Classic5
    th TestRDN_CAR_Y.lua -model RDN_CAR_Y_Q30 -noise 30 -selfEnsemble false -dataset Classic5
    th TestRDN_CAR_Y.lua -model RDN_CAR_Y_Q40 -noise 40 -selfEnsemble false -dataset Classic5
    
    ## With self-ensemble: RDN+
    th TestRDN_CAR_Y.lua -model RDN_CAR_Y_Q10 -noise 10 -selfEnsemble true -dataset Classic5
    th TestRDN_CAR_Y.lua -model RDN_CAR_Y_Q20 -noise 20 -selfEnsemble true -dataset Classic5
    th TestRDN_CAR_Y.lua -model RDN_CAR_Y_Q30 -noise 30 -selfEnsemble true -dataset Classic5
    th TestRDN_CAR_Y.lua -model RDN_CAR_Y_Q40 -noise 40 -selfEnsemble true -dataset Classic5

Results

PSNR_SSIM_BI Table 1. Benchmark results with BI degradation model. Average PSNR/SSIM values for scaling factor ×2, ×3, and ×4.

PSNR_SSIM_BD_DN Table 2. Benchmark results with BD and DN degradation models. Average PSNR/SSIM values for scaling factor ×3.

Citation

If you find the code helpful in your resarch or work, please cite the following papers.

@InProceedings{Lim_2017_CVPR_Workshops,
  author = {Lim, Bee and Son, Sanghyun and Kim, Heewon and Nah, Seungjun and Lee, Kyoung Mu},
  title = {Enhanced Deep Residual Networks for Single Image Super-Resolution},
  booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
  month = {July},
  year = {2017}
}

@inproceedings{zhang2018residual,
    title={Residual Dense Network for Image Super-Resolution},
    author={Zhang, Yulun and Tian, Yapeng and Kong, Yu and Zhong, Bineng and Fu, Yun},
    booktitle={CVPR},
    year={2018}
}

@article{zhang2018rdnir,
    title={Residual Dense Network for Image Restoration},
    author={Zhang, Yulun and Tian, Yapeng and Kong, Yu and Zhong, Bineng and Fu, Yun},
    booktitle={arXiv},
    year={2018}
}

Acknowledgements

This code is built on EDSR (Torch). We thank the authors for sharing their codes of EDSR Torch version and PyTorch version.

You can’t perform that action at this time.