SLRNet: Super Lightweight Residual Network for Real-Time Image Dehazing

1.Overview

• This repo provides a light-weight, end-to-end image dehazing network.

2.Installation

# Python 3.10+ and CUDA 12.4 recommended
pip install -r requirements.txt

3.Usage

1. Specify the training-set and test-set path.

# train & test
python -m SLRNet

4.Citation

This paper has been accepted by ICIC 2026 (Link).

@inproceedings{qu2026slrnet,
  title={SLRNet: Super Lightweight Residual Network for Real-Time Image Dehazing},
  author={Qu, Guanheng and Jiang, Fan and Liu, Jiangming},
  journal={},
  year={2026},
  address={Toronto, Canada},
  month={July},
  url={},
  doi = {10.5281/zenodo.20051973},
  note={Accepted for publication},
}