arxiv:2603.25502

RealRestorer: Towards Generalizable Real-World Image Restoration with Large-Scale Image Editing Models

Published on Mar 26

· Submitted by

Wei Cheng on Mar 27

#3 Paper of the day

Southern university of science and technology

Upvote

Authors:

Yufeng Yang ,

Wei Cheng ,

Abstract

A large-scale dataset and open-source model are developed to improve image restoration performance and close the gap with closed-source alternatives, with a dedicated benchmark for real-world degradation evaluation.

AI-generated summary

Image restoration under real-world degradations is critical for downstream tasks such as autonomous driving and object detection. However, existing restoration models are often limited by the scale and distribution of their training data, resulting in poor generalization to real-world scenarios. Recently, large-scale image editing models have shown strong generalization ability in restoration tasks, especially for closed-source models like Nano Banana Pro, which can restore images while preserving consistency. Nevertheless, achieving such performance with those large universal models requires substantial data and computational costs. To address this issue, we construct a large-scale dataset covering nine common real-world degradation types and train a state-of-the-art open-source model to narrow the gap with closed-source alternatives. Furthermore, we introduce RealIR-Bench, which contains 464 real-world degraded images and tailored evaluation metrics focusing on degradation removal and consistency preservation. Extensive experiments demonstrate our model ranks first among open-source methods, achieving state-of-the-art performance.

View arXiv page View PDF Project page GitHub 39 Add to collection

Community

wchengad

Paper author Paper submitter about 14 hours ago

•

edited about 13 hours ago

RealRestorer: Towards Generalizable Real-World Image Restoration with Large-Scale Image Editing Models

RealRestorer explores how large-scale image editing models can be used for generalizable real-world image restoration.
Instead of focusing on a single degradation type, it aims to handle diverse real-world corruptions within a unified framework while preserving image fidelity and natural visual details.

Highlights

A unified framework for generalizable real-world image restoration
Built upon large-scale image editing models to improve restoration robustness
Designed for diverse degradation types rather than a single specialized setting
Public model release available on Hugging Face

Citation

@misc {yang2026realrestorergeneralizablerealworldimage,
      title={RealRestorer: Towards Generalizable Real-World Image Restoration with Large-Scale Image Editing Models},
      author={Yufeng Yang and Xianfang Zeng and Zhangqi Jiang and Fukun Yin and Jianzhuang Liu and Wei Cheng and jinghong lan and Shiyu Liu and Yuqi Peng and Gang YU and Shifeng Chen},
      year={2026},
      eprint={2603.25502},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2603.25502},
}