bruce-wanggf/msffdn

ACCV2022 MSFFDN for dehazing

MSF$^2$DN: Mutil Scale Feature Fusion Dehazing Network with Dense Connection (ACCV2022)

This is the official PyTorch codes for the paper.

MSF$^2$DN: Mutil Scale Feature Fusion Dehazing Network with Dense Connection
Guangfa Wang, Xiaokang Yu* (* indicates corresponding author)
The 16th Asian Conference on Computer Vision， ACCV2022

Demo

Get Started

Prepare models

Downloading checkpoints

Model	Description	🔗 Download Links
MSF$^2$DN	Mutil Scale Feature Fusion Dehazing Network with Dense Connection	[Baidu Disk (pwd: 5z34)]

Quick demo

Run demos to process the images in dir ./text_image/ by following command, before this, you should input the right path in dehaze.py:

python dehaze.py

Train MSF$^2$DN

Train our MSF$^2$DN, before this, you should input the right path in the corresponding files. In the dataloader.py, you should configure the right name of picture and then run the following command:

python train.py

Citation

If you find our repo useful for your research, please cite us:

@InProceedings{Wang2022ACCV,
    author    = {Wang, Guangfa and Yu, Xiaokang},
    title     = {MSF\${\textasciicircum}2\$DN:Multi Scale Feature Fusion Dehazing  Network with Dense connection},
    booktitle = {Proceedings of the Asian Conference on Computer Vision (ACCV)},
    month     = {December},
    year      = {2022},
    pages     = {2950-2966}
}

Acknowledgement

This repository is maintained by Guangfa Wang.

msffdn's Issues

资源失效

谢谢作者的分享，但是云盘失效了，请问能再次分享一下吗，十分感谢！

great job！Your paper on dehazing is outstanding, and your research has made significant contributions to the field~