[ECCV 2024] Restoring Images in Adverse Weather Conditions via Histogram Transformer

Cover figure	Network structure

---

Restoring Images in Adverse Weather Conditions via Histogram Transformer
Shangquan Sun, Wenqi Ren, Xinwei Gao, Rui Wang, Xiaochun Cao
European Conference on Computer Vision

Abstract: Transformer-based image restoration methods in adverse weather have achieved significant progress. Most of them use self-attention along the channel dimension or within spatially fixed-range blocks to reduce computational load. However, such a compromise results in limitations in capturing long-range spatial features. Inspired by the observation that the weather-induced degradation factors mainly cause similar occlusion and brightness, in this work, we propose an efficient Histogram Transformer (Histoformer) for restoring images affected by adverse weather. It is powered by a mechanism dubbed histogram self-attention, which sorts and segments spatial features into intensity-based bins. Self-attention is then applied across bins or within each bin to selectively focus on spatial features of dynamic range and process similar degraded pixels of the long range together. To boost histogram self-attention, we present a dynamic-range convolution enabling conventional convolution to conduct operation over similar pixels rather than neighbor pixels. We also observe that the common pixel-wise losses neglect linear association and correlation between output and ground-truth. Thus, we propose to leverage the Pearson correlation coefficient as a loss function to enforce the recovered pixels following the identical order as ground-truth. Extensive experiments demonstrate the efficacy and superiority of our proposed method. We have released the codes in Github

---

🚀 News

• 2024.07.23: Update alternative [Download Links] for Snow100K as its original link expires.
• 2024.07.18: Codes and pre-trained weights are released.
• 2024.07.17: Visual results are released.
• 2024.07.14: Arxiv Paper is released.
• 2024.07.01: Histoformer is accepted by ECCV2024.
• 2024.03.01: Histoformer is rejected by CVPR2024.

😄 Visual Results

All visual results are in Google Drive and Baidu Disk (pin: ps9q). You can also find each of them from the table below.

Method	RainDrop	Outdoor-Rain	Snow100K-L	Snow100K-S	RealSnow
TransWeather	Results	Results	Results	Results	Results
WGWSNet	Results	Results	Results	Results	Results
Chen et al.	Results	Results	Results	Results	Results
WeatherDiff	Results	Results	Results	Results	Results
Histoformer	Results	Results	Results	Results	Results

Examples:

RainDrop	Outdoor-Rain

Snow100K-L	RealSnow

⚙️ Installation

See INSTALL.md for the installation of dependencies required to run Histoformer.

🛠️ Training

1. Download Training set or each of them, i.e., Snow100K, Outdoor-Rain, and RainDrop.

Note: The original link for downloading Snow100K has expired, so you could refer to [Issue#2] for alternative download links.

2. Modify the configurations of dataroot_gt and dataroot_lq for train, val_snow_s, val_snow_l, val_test1 and val_raindrop in Allweather/Options/Allweather_Histoformer.yml

3. To train Histoformer with default settings, run

cd Histoformer
./train.sh Allweather/Options/Allweather_Histoformer.yml 4321

Note: The above training script uses 4 GPUs by default. To use any other settings, modify Histoformer/train.sh and Allweather/Options/Allweather_Histoformer.yml

⚖️ Evaluation

0. cd Allweather

1. Download the pre-trained models and place it in ./pretrained_models/

2. Download test datasets from each of them, i.e., Snow100K, Outdoor-Rain, and RainDrop.

3. Testing with the replaced argument --input_dir [INPUT_FOLDER]

python test_histoformer.py --input_dir [INPUT_FOLDER] --result_dir result/ --weights pretrained_models/net_g_best.pth --yaml_file Options/Allweather_Histoformer.yml

# for realsnow
python test_histoformer.py --input_dir [INPUT_FOLDER] --result_dir result/ --weights pretrained_models/net_g_real.pth --yaml_file Options/Allweather_Histoformer.yml

⚖️ Demo

0. cd Allweather

1. Download the pre-trained models and place it in ./pretrained_models/

2. Testing with the replaced argument --input_dir [INPUT_FOLDER]

# for realsnow
python test_histoformer.py --input_dir [INPUT_FOLDER] --result_dir result/ --weights pretrained_models/net_g_real.pth --yaml_file Options/Allweather_Histoformer.yml

📣 Citation

If you use Histoformer, please consider citing:

@misc{sun2024restoring,
title={Restoring Images in Adverse Weather Conditions via Histogram Transformer},
author={Shangquan Sun and Wenqi Ren and Xinwei Gao and Rui Wang and Xiaochun Cao},
year={2024},
eprint={2407.10172},
archivePrefix={arXiv},
primaryClass={cs.CV}
}

📬 Contact

If you have any question, please contact [email protected]

Acknowledgment: This code is based on the Restormer, BasicSR toolbox and HINet.