Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/dhryougit/AFM
https://github.com/dhryougit/AFM
Last synced: 27 days ago
JSON representation
- Host: GitHub
- URL: https://github.com/dhryougit/AFM
- Owner: dhryougit
- License: other
- Created: 2024-03-28T12:48:27.000Z (8 months ago)
- Default Branch: main
- Last Pushed: 2024-06-18T18:17:08.000Z (5 months ago)
- Last Synced: 2024-07-13T06:22:46.888Z (4 months ago)
- Language: Python
- Size: 563 KB
- Stars: 15
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: readme.md
- License: LICENSE
Awesome Lists containing this project
- Awesome-Mixup - [Code
README
# Robust Image Denoising through Adversarial Frequency Mixup (2024 CVPR) [[paper]](https://openaccess.thecvf.com/content/CVPR2024/html/Ryou_Robust_Image_Denoising_through_Adversarial_Frequency_Mixup_CVPR_2024_paper.html)
### Installation
This implementation based on [BasicSR](https://github.com/xinntao/BasicSR) which is a open source toolbox for image/video restoration tasks and [NAFNet](https://github.com/megvii-research/NAFNet)
```python
python 3.8.8
pytorch 1.9.0
cuda 11.3
```
```
git clone https://github.com/dhryougit/AFM.git
cd AFM
pip install -r requirements.txt
python setup.py develop --no_cuda_ext
```
We used NVIDIA RTX A6000 D6 48GB for trianing our models.
### QuickStart
For training
```
python3 -m torch.distributed.launch --nproc_per_node=4 --master_port=4321 basicsr/train.py -opt options/train/DnCNN.yml --name=DnCNN-afm-b --afm_type=AFM_B --seed=10 --afm_rate=0.8 --afm_easy_rate=0.3 --launcher pytorch
```
### Dataset
Training dataset : [SIDD](https://abdokamel.github.io/sidd/#sidd-medium)
Evaludation datasets : [Poly](https://github.com/csjunxu/PolyU-Real-World-Noisy-Images-Dataset), [CC](https://github.com/csjunxu/MCWNNM-ICCV2017), HighISO, iPhone, Huawei.
Additioanl real-world noise datasets can be downloaded from "https://github.com/ZhaomingKong/Denoising-Comparison"
### Results and Pre-trained model
| Dataset | Poly |CC |HighISO |iPhone |Huawei | OOD Avg.|
|:----|:----|:----|:----|:----|:----|-----|
|PSRN| 37.75 | 36.84 | 39.17 | 40.65 | 38.39 | 38.56 |
|SSIM| 0.9804 | 0.9830 | 0.9801 | 0.9777 | 0.9683 | 0.9779 |
Pre-trained model of our Dncnn trained on AFM-B can be downloaded from (https://drive.google.com/file/d/1uPJP2zNc4ViFc1QU7TXGAwrFIEScBJvL/view?usp=sharing)