Authors: Yinglong Wang, Bin He
Paper link(arxiv): Link of paper
The project is built with Python 3.8, PyTorch 1.10.0, CUDA 11.3,
For installing, follow these instructions:
pip install -r requirements.txt
If you're having problems with your version, you can try installing each package one by one using pip alone
To Train ,you can run this cmd in your terminal:
python main.py --mode train --dataset yourpath --batch_size 8
You can also change the parameters in the main.py file.Saving your changes and running main.py will have the same effect
You can download the pre-trained models from this link:
Download the pre-trained models
If you just need the output images of the models, you can download the images from this link:
Dataset | PSNR | SSIM |
---|---|---|
SOTS-Indoor | 43.21 | 0.997 |
SOTS-Outdoor | 38.86 | 0.995 |
Dense-Haze | 17.56 | 0.658 |
O-HAZE | 25.44 | 0.936 |
Haze4K | 35.73 | 0.99 |
Will be provided when the paper is accepted.
name: Yinglong Wang
e-mail: wangyinglong2023@gmail.com
wechat: dauing2023