RIRF: Reasoning Image Restoration Framework

2026-04-10

AI Summary

Existing general-purpose image restoration methods lack explicit diagnostic reasoning about degradation types, severity levels, and scene semantics, which limits their restoration performance. This work proposes the Reason and Restore (R&R) framework, which for the first time tightly integrates structured chain-of-thought reasoning with pixel-level restoration within a unified architecture. Specifically, a fine-tuned Qwen3-VL model serves as an interpretable reasoning module that explicitly diagnoses degradation characteristics and generates fine-grained priors. Furthermore, degradation severity is leveraged as a reinforcement learning signal to enable end-to-end joint optimization of reasoning and restoration. The proposed method achieves state-of-the-art performance across multiple general image restoration benchmarks while offering an interpretable diagnostic process for the restoration pipeline.

Technology Category

• Computer Vision: Visual Reasoning & Symbolic Representations
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Universal image restoration (UIR) aims to recover clean images from diverse and unknown degradations using a unified model. Existing UIR methods primarily focus on pixel reconstruction and often lack explicit diagnostic reasoning over degradation composition, severity, and scene semantics prior to restoration. We propose Reason and Restore (R\&R), a novel framework that integrates structured Chain-of-Thought (CoT) reasoning into the image restoration pipeline. R\&R introduces an explicit reasoner, implemented by fine-tuning Qwen3-VL, to diagnose degradation types, quantify degradation severity, infer key degradation-related factors, and describe relevant scene and object semantics. The resulting structured reasoning provides interpretable and fine-grained diagnostic priors for the restorer. To further improve restoration quality, the quantified degradation severity produced by the reasoner is leveraged as reinforcement learning (RL) signals to guide and strengthen the restorer. Unlike existing multimodal LLM-based agentic systems that decouple reasoning from low-level vision tasks, R\&R tightly couples semantic diagnostic reasoning with pixel-level restoration in a unified framework. Extensive experiments across diverse UIR benchmarks demonstrate that R\&R achieves state-of-the-art performance while offering unique interpretability into the restoration process.

Problem

• universal image restoration
• degradation reasoning
• scene semantics
• image degradation
• diagnostic reasoning

Innovation

• Chain-of-Thought Reasoning
• Universal Image Restoration
• Multimodal Reasoning
• Reinforcement Learning
• Interpretable Restoration