This repository is implementation of the "Accelerating the Super-Resolution Convolutional Neural Network".
- Added the zero-padding
- Used the Adam instead of the SGD
- PyTorch 1.0.0
- Numpy 1.15.4
- Pillow 5.4.1
- h5py 2.8.0
- tqdm 4.30.0
The 91-image, Set5 dataset converted to HDF5 can be downloaded from the links below.
| Dataset | Scale | Type | Link |
|---|---|---|---|
| 91-image | 2 | Train | Download |
| 91-image | 3 | Train | Download |
| 91-image | 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/91-image_x3.h5" \
--eval-file "BLAH_BLAH/Set5_x3.h5" \
--outputs-dir "BLAH_BLAH/outputs" \
--scale 3 \
--lr 1e-3 \
--batch-size 16 \
--num-epochs 20 \
--num-workers 8 \
--seed 123 Pre-trained weights can be downloaded from the links below.
| Model | Scale | Link |
|---|---|---|
| FSRCNN(56,12,4) | 2 | Download |
| FSRCNN(56,12,4) | 3 | Download |
| FSRCNN(56,12,4) | 4 | Download |
The results are stored in the same path as the query image.
python test.py --weights-file "BLAH_BLAH/fsrcnn_x3.pth" \
--image-file "data/butterfly_GT.bmp" \
--scale 3PSNR was calculated on the Y channel.
| Eval. Mat | Scale | Paper | Ours (91-image) |
|---|---|---|---|
| PSNR | 2 | 36.94 | 37.12 |
| PSNR | 3 | 33.06 | 33.22 |
| PSNR | 4 | 30.55 | 30.50 |
| Original | BICUBIC x3 | FSRCNN x3 (34.66 dB) |
|
|
|
| Original | BICUBIC x3 | FSRCNN x3 (28.55 dB) |
|
|
|
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent