Project

Residual Diffusion Bridge Model for Image Restoration

Hebaixu Wang, Jing Zhang, Haoyang Chen, Haonan Guo, Di Wang, Jiayi Ma and Bo Du.

Paper | Github Code

Abstract

Diffusion bridge models establish probabilistic paths between arbitrary paired distributions and exhibit great potential for universal image restoration. Most existing methods merely treat them as simple variants of stochastic interpolants, lacking a unified analytical perspective. Besides, they indiscriminately reconstruct images through global noise injection and removal, inevitably distorting undegraded regions due to imperfect reconstruction. To address these challenges, we propose the {R}esidual {D}iffusion {B}ridge {M}odel (RDBM). Specifically, we theoretically reformulate the stochastic differential equations of generalized diffusion bridge and derive the analytical formulas of its forward and reverse processes. Crucially, we leverage the residuals from given distributions to modulate the noise injection and removal, enabling adaptive restoration of degraded regions while preserving intact others. Additionally, we unravel the fundamental mathematical essence of existing bridge models, all of which are special cases of RDBM and empirically demonstrate the optimality of our proposed models. Extensive experiments are conducted to demonstrate the state-of-the-art performance of our method both qualitatively and quantitatively across diverse image restoration tasks.

Introducation

Overview

Stochastic Trajectories

Visualization

Datasets Information

Task	Dataset	Synthetic/Real	Download Links
Deraining	DID	Synthetic	URL
	DeRaindrop	Real	URL
	Rain13K	Synthetic	URL
	Rain_100H	Synthetic	URL
	Rain_100L	Synthetic	URL
	GT-Rain	Real	URL
	RealRain-1k	Real	URL
Low-light Enhancement	LOL	Real	URL
	MEF	Real	URL
	VE-LOL-L	Synthetic/Real	URL
	NPE	Real	URL
Desnowing	CSD	Synthetic	URL
	Snow100K-Real	Real	URL
Dehazing	SOTS	Synthetic	URL
	ITS_v2	Synthetic	URL
	D-HAZY	Synthetic	URL
	NH-HAZE	Real	URL
	Dense-Haze	Real	URL
	NHRW	Real	URL
Deblur	GoPro	Synthetic	URL
	RealBlur	Real	URL

Contributor

Baixuzx7 @ wanghebaixu@gmail.com

Citation

@inproceedings{wang2026residual,
  title={Residual diffusion bridge model for image restoration},
  author={Wang, Hebaixu and Zhang, Jing and Chen, Haoyang and Guo, Haonan and Wang, Di and Ma, Jiayi and Du, Bo},
  booktitle={Proceedings of the Conference on Computer Vision and Pattern Recognition},
  pages={8375--8386},
  year={2026}
}

Copyright statement

The project is signed under the MIT license, see the LICENSE.md