https://github.com/hejingwenhejingwen/adafm
CVPR2019 (oral) Modulating Image Restoration with Continual Levels via Adaptive Feature Modification Layers (AdaFM). PyTorch implementation
https://github.com/hejingwenhejingwen/adafm
adafm cvpr2019 cvpr2019-oral denoising image-restoration low-level-vision pytorch super-resolution
Last synced: 4 months ago
JSON representation
CVPR2019 (oral) Modulating Image Restoration with Continual Levels via Adaptive Feature Modification Layers (AdaFM). PyTorch implementation
- Host: GitHub
- URL: https://github.com/hejingwenhejingwen/adafm
- Owner: hejingwenhejingwen
- Created: 2019-04-29T15:41:28.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2019-08-05T13:51:58.000Z (about 6 years ago)
- Last Synced: 2025-04-06T12:40:05.301Z (6 months ago)
- Topics: adafm, cvpr2019, cvpr2019-oral, denoising, image-restoration, low-level-vision, pytorch, super-resolution
- Language: Python
- Homepage: https://arxiv.org/abs/1904.08118
- Size: 7.37 MB
- Stars: 184
- Watchers: 4
- Forks: 21
- Open Issues: 7
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Modulating Image Restoration with Continual Levels via Adaptive Feature Modification Layers [paper](https://arxiv.org/abs/1904.08118), [supplementary file](http://openaccess.thecvf.com/content_CVPR_2019/supplemental/He_Modulating_Image_Restoration_CVPR_2019_supplemental.pdf)
By Jingwen He, [Chao Dong](https://scholar.google.com.hk/citations?user=OSDCB0UAAAAJ&hl=en), and [Yu Qiao](http://mmlab.siat.ac.cn/yuqiao/)
```python
class AdaptiveFM(nn.Module):
def __init__(self, in_channel, kernel_size):
super(AdaptiveFM, self).__init__()
padding = (kernel_size - 1) // 2
self.transformer = nn.Conv2d(in_channel, in_channel, kernel_size, padding=padding, groups=in_channel)
def forward(self, x):
return self.transformer(x) + x
```
### BibTex
@InProceedings{He_2019_CVPR,
author = {He, Jingwen and Dong, Chao and Qiao, Yu},
title = {Modulating Image Restoration With Continual Levels via Adaptive Feature Modification Layers},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2019}
}
### Dependencies
- Python 3 (Recommend to use [Anaconda](https://www.anaconda.com/download/#linux))
- [PyTorch >= 0.4.0](https://pytorch.org/)
- NVIDIA GPU + [CUDA](https://developer.nvidia.com/cuda-downloads)
- Python packages: `pip install numpy opencv-python lmdb`
- [option] Python packages: [`pip install tensorboardX`](https://github.com/lanpa/tensorboardX), for visualizing curves.
# Pretrained models
We provide a pretrained model for AdaFM-Net ([`experiments/pretrained_models`](experiments/pretrained_models)) that deals with denoising from σ15 to σ75. Please run the following commands directly:
```c++
cd codes
python interpolate.py -opt options/test/test.json
```
The results can be found in the newly created directory `AdaFM/results`
The noise level of the [`input image`](datasets/personal_images/personal_images_noise45/soilder.png) is σ45, and you are supposed to obtain similar interpolated results as follows:
# Codes
The overall code framework mainly consists of four parts - Config, Data, Model and Network.
We also provides some useful scripts.
Please run all the following commands in “codes” directory.
## How to Test
### basic model and AdaFM-Net
1. Modify the configuration file [`options/test/test.json`](codes/options/test/test.json) (please refer to [`options`](codes/options) for instructions.)
1. Run command:
```c++
python test.py -opt options/test/test.json
```
### modulation testing
1. Modify the configuration file [`options/test/test.json`](codes/options/test/test.json)
1. Run command:
```c++
python interpolate.py -opt options/test/test.json
```
#### or:
1. Use [`scripts/net_interp.py`](codes/scripts/net_interp.py) to obtain the interpolated network.
1. Modify the configuration file [`options/test/test.json`](codes/options/test/test.json) and run command: `python test.py -opt options/test/test.json`
## How to Train
### basic model
1. Prepare datasets, usually the DIV2K dataset. More details are in [`codes/data`](codes/data).
1. Modify the configuration file [`options/train/train_basic.json`](codes/options/train/train_basic.json) (please refer to [`options`](codes/options) for instructions.)
1. Run command:
```c++
python train.py -opt options/train/train_basic.json
```
### AdaFM-Net
1. Prepare datasets, usually the DIV2K dataset.
1. Modify the configuration file [`options/train/train_adafm.json`](codes/options/train/train_adafm.json)
1. Run command:
```c++
python train.py -opt options/train/train_adafm.json
```
## Acknowledgement
- This code borrows heavily from [BasicSR](https://github.com/xinntao/BasicSR).