QuReC: All-in-One Image Restoration with Query-Specific Guidance and Local-Global Response Calibration

2026-07-16

AI Summary

Existing single models struggle to effectively handle the spatially heterogeneous and coexisting mixture of degradations commonly present in real-world images. To address this challenge, this work proposes QuReC, a unified image restoration framework featuring two core innovations: a degradation prototype space that enables query-level adaptive guidance, and a local-global dual-branch response calibration mechanism designed to enhance the reliability of feature aggregation. Furthermore, the framework integrates weakly supervised prototype matching with a learnable prior calibration strategy to better capture degradation characteristics. Extensive experiments demonstrate that QuReC significantly outperforms current state-of-the-art methods across multiple universal image restoration benchmarks, achieving leading performance in both quantitative metrics and visual quality.

Abstract

All-in-one image restoration aims to recover clean images degraded by multiple corruption types using a single unified model. Existing methods typically rely on image-level prompts or shared guidance to handle diverse degradations. However, such a paradigm becomes inadequate when degradations are spatially heterogeneous or even coexist in mixed forms within a single image. Yet spatially adaptive guidance alone is not sufficient, since accurate restoration also requires each spatial query to reliably aggregate complementary information from local neighborhoods and global contexts. To this end, we propose QuReC, a unified framework for all-in-one image restoration. QuReC consists of a Degradation-Guided Query Reconstruction Module (DQRM) and a Local-Global Response Calibration Module (LGRCM). Specifically, DQRM matches each spatial query against a degradation prototype space to reconstruct a query-specific degradation-aware representation, thereby providing fine-grained spatially adaptive restoration guidance. To further stabilize this query-wise matching process, we introduce a weakly supervised prototype matching learning strategy to improve optimization stability and degradation semantic consistency. Meanwhile, LGRCM performs local-global dual-branch aggregation and calibrates the aggregated responses with learnable priors, improving the reliability of feature aggregation and the coordination between local detail modeling and global context modeling. Extensive experiments demonstrate that QuReC achieves superior performance on multiple all-in-one image restoration benchmarks. The code is released at https://github.com/zhoushen1/QuReC.

Problem

• all-in-one image restoration
• spatially heterogeneous degradations
• mixed degradations
• local-global feature aggregation
• degradation-aware guidance

Innovation

• query-specific guidance
• local-global calibration
• degradation prototype
• all-in-one image restoration
• spatially adaptive restoration