RDN
This repository is implementation of the "Residual Dense Network for Image Super-Resolution".
Requirements
- PyTorch 1.0.0
- Numpy 1.15.4
- Pillow 5.4.1
- h5py 2.8.0
- tqdm 4.30.0
Train
The DIV2K, Set5 dataset converted to HDF5 can be downloaded from the links below.
| Dataset | Scale | Type | Link |
|---|---|---|---|
| DIV2K | 2 | Train | Download |
| DIV2K | 3 | Train | Download |
| DIV2K | 4 | Train | Download |
| Set5 | 2 | Eval | Download |
| Set5 | 3 | Eval | Download |
| Set5 | 4 | Eval | Download |
Otherwise, you can use prepare.py to create custom dataset.
python train.py --train-file "BLAH_BLAH/DIV2K_x4.h5" \
--eval-file "BLAH_BLAH/Set5_x4.h5" \
--outputs-dir "BLAH_BLAH/outputs" \
--scale 4 \
--num-features 64 \
--growth-rate 64 \
--num-blocks 16 \
--num-layers 8 \
--lr 1e-4 \
--batch-size 16 \
--patch-size 32 \
--num-epochs 800 \
--num-workers 8 \
--seed 123 Test
Pre-trained weights can be downloaded from the links below.
| Model | Scale | Link |
|---|---|---|
| RDN (D=16, C=8, G=64, G0=64) | 2 | Download |
| RDN (D=16, C=8, G=64, G0=64) | 3 | Download |
| RDN (D=16, C=8, G=64, G0=64) | 4 | Download |
The results are stored in the same path as the query image.
python test.py --weights-file "BLAH_BLAH/rdn_x4.pth" \
--image-file "data/119082.png" \
--scale 4 \
--num-features 64 \
--growth-rate 64 \
--num-blocks 16 \
--num-layers 8
Results
PSNR was calculated on the Y channel.
Set5
| Eval. Mat | Scale | RDN (Paper) | RDN (Ours) |
|---|---|---|---|
| PSNR | 2 | 38.24 | 38.18 |
| PSNR | 3 | 34.71 | 34.73 |
| PSNR | 4 | 32.47 | 32.40 |
| Original | BICUBIC x4 | RDN x4 (25.08 dB) |
|
|
|
| Original | BICUBIC x4 | RDN x4 (32.98 dB) |
|
|
|
Citation
@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{zhang2020rdnir,
title={Residual Dense Network for Image Restoration},
author={Zhang, Yulun and Tian, Yapeng and Kong, Yu and Zhong, Bineng and Fu, Yun},
journal={TPAMI},
year={2020}
}