MADN

https://github.com/dehazing/MADN

Effective Meta-Attention Dehazing Networks for Vision-Based Outdoor Industrial Systems

Prerequisites

　 Python3.6

　 PyTorch>=1.0.1

　 torchvision>=0.4.2

　 skimage

　 tqdm

Quick Start

Testing:

Clone this repo in environment that satisfies the prerequisites

Run test.py using default hyper-parameter settings.

(You can test on the trained model: dehaze_80.pth is the pre-trained meta network, model.pth is the pre-trained dehazing network.)

Training:

Run main.py for training.

Cite

If you use part of this code, please kindly cite

 @article{jia2021effective,
  title={Effective Meta-Attention Dehazing Networks for Vision-Based Outdoor Industrial Systems},
  author={Jia, Tongyao and Li, Jiafeng and Zhuo, Li and Li, Guoqiang},
  journal={IEEE Transactions on Industrial Informatics},
  year={2021},
  publisher={IEEE}
}