IRBridge: Solving Image Restoration Bridge with Pre-trained Generative Diffusion Models

Hanting Wang, Tao Jin, Wang Lin, Shulei Wang, Hai Huang, Shengpeng Ji, Zhou Zhao
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:63266-63289, 2025.

Abstract

Bridge models in image restoration construct a diffusion process from degraded to clear images. However, existing methods typically require training a bridge model from scratch for each specific type of degradation, resulting in high computational costs and limited performance. This work aims to efficiently leverage pretrained generative priors within existing image restoration bridges to eliminate this requirement. The main challenge is that standard generative models are typically designed for a diffusion process that starts from pure noise, while restoration tasks begin with a low-quality image, resulting in a mismatch in the state distributions between the two processes. To address this challenge, we propose a transition equation that bridges two diffusion processes with the same endpoint distribution. Based on this, we introduce the IRBridge framework, which enables the direct utilization of generative models within image restoration bridges, offering a more flexible and adaptable approach to image restoration. Extensive experiments on six image restoration tasks demonstrate that IRBridge efficiently integrates generative priors, resulting in improved robustness and generalization performance. Code will be available at GitHub.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-wang25as, title = {{IRB}ridge: Solving Image Restoration Bridge with Pre-trained Generative Diffusion Models}, author = {Wang, Hanting and Jin, Tao and Lin, Wang and Wang, Shulei and Huang, Hai and Ji, Shengpeng and Zhao, Zhou}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {63266--63289}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/wang25as/wang25as.pdf}, url = {https://proceedings.mlr.press/v267/wang25as.html}, abstract = {Bridge models in image restoration construct a diffusion process from degraded to clear images. However, existing methods typically require training a bridge model from scratch for each specific type of degradation, resulting in high computational costs and limited performance. This work aims to efficiently leverage pretrained generative priors within existing image restoration bridges to eliminate this requirement. The main challenge is that standard generative models are typically designed for a diffusion process that starts from pure noise, while restoration tasks begin with a low-quality image, resulting in a mismatch in the state distributions between the two processes. To address this challenge, we propose a transition equation that bridges two diffusion processes with the same endpoint distribution. Based on this, we introduce the IRBridge framework, which enables the direct utilization of generative models within image restoration bridges, offering a more flexible and adaptable approach to image restoration. Extensive experiments on six image restoration tasks demonstrate that IRBridge efficiently integrates generative priors, resulting in improved robustness and generalization performance. Code will be available at GitHub.} }
Endnote
%0 Conference Paper %T IRBridge: Solving Image Restoration Bridge with Pre-trained Generative Diffusion Models %A Hanting Wang %A Tao Jin %A Wang Lin %A Shulei Wang %A Hai Huang %A Shengpeng Ji %A Zhou Zhao %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-wang25as %I PMLR %P 63266--63289 %U https://proceedings.mlr.press/v267/wang25as.html %V 267 %X Bridge models in image restoration construct a diffusion process from degraded to clear images. However, existing methods typically require training a bridge model from scratch for each specific type of degradation, resulting in high computational costs and limited performance. This work aims to efficiently leverage pretrained generative priors within existing image restoration bridges to eliminate this requirement. The main challenge is that standard generative models are typically designed for a diffusion process that starts from pure noise, while restoration tasks begin with a low-quality image, resulting in a mismatch in the state distributions between the two processes. To address this challenge, we propose a transition equation that bridges two diffusion processes with the same endpoint distribution. Based on this, we introduce the IRBridge framework, which enables the direct utilization of generative models within image restoration bridges, offering a more flexible and adaptable approach to image restoration. Extensive experiments on six image restoration tasks demonstrate that IRBridge efficiently integrates generative priors, resulting in improved robustness and generalization performance. Code will be available at GitHub.
APA
Wang, H., Jin, T., Lin, W., Wang, S., Huang, H., Ji, S. & Zhao, Z.. (2025). IRBridge: Solving Image Restoration Bridge with Pre-trained Generative Diffusion Models. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:63266-63289 Available from https://proceedings.mlr.press/v267/wang25as.html.

Related Material