Introduction

This is the official PyTorch implementation of DehazeDCT: Towards Effective Non-Homogeneous Dehazing via Deformable Convolutional Transformer in CVPRW 2024. Our DehazeDCT won the second place in the NTIRE 2024 Dense and Non-Homogeneous Dehazing Challenge. If you find this repo useful, please give it a star ⭐ and consider citing our paper in your research. Thank you.

NTIRE 2024 Award

 

 

Overall Framework

Results

Performance on NH-HAZE, NH-HAZE2, HD-NH-HAZE, and DNH-HAZE:

Rank of NTIRE 2024 Dense and Non-Homogeneous Dehazing Challenge:

Result of DehazeDCT for NTIRE 2024 Dense and Non-Homogeneous Dehazing Challenge Test Data:

Create Environment

Dependencies and Installation

• Python 3.8
• Pytorch 1.11

1. Create Conda Environment

conda create --name dehazedct python=3.8
conda activate dehazedct

2. Install Dependencies

conda install pytorch=1.11 torchvision cudatoolkit=11.3 -c pytorch
pip install -U openmim
mim install mmcv-full==1.5.0
pip install timm==0.6.11 mmdet==2.28.1
pip install opencv-python termcolor yacs pyyaml scipy
pip install DCNv4==latest
pip install numpy matplotlib scikit-learn scikit-image timm kornia einops pytorch_lightning

Datasets

NH-HAZE NTIRE 2020

NH-HAZE2 NTIRE 2021

HD-NH-HAZE NTIRE 2023

DNH-HAZE NTIRE 2024

Pre-trained Model

• NTIRE 2024 Dense and Non-Homogeneous Dehazing Challenge.

Testing

Download above saved models and unzip it into the folder ./weights. To test the model, you need to specify the test dictionary (Line 17 in test.py) and model path (Line 41 in test.py and Line 19 in model.py). Then run

python test.py 

You can check the output in ../results.

Citation

If you find our work useful for your research, please consider citing our paper

@InProceedings{dehazedct_2024_cvprw,
    author    = {Dong, Wei and Zhou, Han and Wang, Ruiyi and Liu, Xiaohong and Zhai, Guangtao and Chen, Jun},
    title     = {DehazeDCT: Towards Effective Non-Homogeneous Dehazing via Deformable Convolutional Transformer},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
    month     = {June},
    year      = {2024},
    pages     = {6405-6414}
}

Contact

If you have any question, please feel free to contact us via [email protected].