GenHaze: Pioneering Controllable One-Step Realistic Haze Generation for Real-World Dehazing
Abstract
Real-world image dehazing is crucial for enhancing visual quality in computer vision applications. However, existing physics-based haze generation paradigms struggle to model the complexities of real-world haze and lack controllability, limiting the performance of existing baselines on real-world images. In this paper, we introduce GenHaze, a pioneering haze generation framework that enables the one-step generation of high-quality, reference-controllable hazy images. GenHaze leverages the pre-trained latent diffusion model (LDM) with a carefully designed clean-to-haze generation protocol to produce realistic hazy images. Additionally, by leveraging its fast, controllable generation of paired high-quality hazy images, we illustrate that existing dehazing baselines can be unleashed in a simple and efficient manner. Extensive experiments indicate that GenHaze achieves visually convincing and quantitatively superior hazy images. It also significantly improves multiple existing dehazing models across 7 non-reference metrics with minimal fine-tuning epochs. Our work demonstrates that LDM possesses the potential to generate realistic degradations, providing an effective alternative to prior generation pipelines.
Cite
Text
Chen et al. "GenHaze: Pioneering Controllable One-Step Realistic Haze Generation for Real-World Dehazing." International Conference on Computer Vision, 2025.Markdown
[Chen et al. "GenHaze: Pioneering Controllable One-Step Realistic Haze Generation for Real-World Dehazing." International Conference on Computer Vision, 2025.](https://mlanthology.org/iccv/2025/chen2025iccv-genhaze/)BibTeX
@inproceedings{chen2025iccv-genhaze,
title = {{GenHaze: Pioneering Controllable One-Step Realistic Haze Generation for Real-World Dehazing}},
author = {Chen, Sixiang and Ye, Tian and Lin, Yunlong and Jin, Yeying and Yang, Yijun and Chen, Haoyu and Lai, Jianyu and Fei, Song and Xing, Zhaohu and Tsung, Fugee and Zhu, Lei},
booktitle = {International Conference on Computer Vision},
year = {2025},
pages = {9194-9205},
url = {https://mlanthology.org/iccv/2025/chen2025iccv-genhaze/}
}