TPGDiff: Hierarchical Triple-Prior Guided Diffusion for Image Restoration

2026-01-28

AI Summary

Existing unified image restoration methods exhibit limited capability in reconstructing severely degraded regions, and the early incorporation of semantic information often compromises spatial structure fidelity. To address these limitations, this work proposes TPGDiff, which introduces a novel hierarchical triple-prior guidance mechanism. Specifically, degradation-aware priors are integrated throughout the entire diffusion process, multi-source structural priors are injected in shallow layers to preserve fine details, and distilled semantic priors are fused in deep layers to enhance content generation. The proposed method achieves state-of-the-art performance on both single- and multi-degradation benchmarks, significantly improving restoration quality and generalization, with particularly remarkable gains in scenarios involving severe degradation.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

All-in-one image restoration aims to address diverse degradation types using a single unified model. Existing methods typically rely on degradation priors to guide restoration, yet often struggle to reconstruct content in severely degraded regions. Although recent works leverage semantic information to facilitate content generation, integrating it into the shallow layers of diffusion models often disrupts spatial structures (\emph{e.g.}, blurring artifacts). To address this issue, we propose a Triple-Prior Guided Diffusion (TPGDiff) network for unified image restoration. TPGDiff incorporates degradation priors throughout the diffusion trajectory, while introducing structural priors into shallow layers and semantic priors into deep layers, enabling hierarchical and complementary prior guidance for image reconstruction. Specifically, we leverage multi-source structural cues as structural priors to capture fine-grained details and guide shallow layers representations. To complement this design, we further develop a distillation-driven semantic extractor that yields robust semantic priors, ensuring reliable high-level guidance at deep layers even under severe degradations. Furthermore, a degradation extractor is employed to learn degradation-aware priors, enabling stage-adaptive control of the diffusion process across all timesteps. Extensive experiments on both single- and multi-degradation benchmarks demonstrate that TPGDiff achieves superior performance and generalization across diverse restoration scenarios. Our project page is: https://leoyjtu.github.io/tpgdiff-project.

Problem

• image restoration
• degradation priors
• semantic priors
• structural priors
• diffusion models

Innovation

• Triple-Prior Guidance
• Hierarchical Diffusion
• Structural Prior
• Semantic Prior
• Degradation-Aware Control