Learning to Restore Multi-Degraded Images via Ingredient Decoupling and Task-Aware Path Adaptation

2025-11-07

AI Summary

Real-world images often suffer from multiple concurrent degradations—such as rain, haze, and noise—whereas existing methods typically address only single degradations, limiting their generalizability. To tackle this, we propose IMDNet, the first framework enabling statistical disentanglement of mixed degradation components and task-adaptive selection of restoration pathways. Specifically, we design a Degradation-Ingredient Disentanglement Block (DIDBlock) to separate entangled degradation features; introduce a Learnable Fusion Block (FBlock) and a Task-Adaptive Block (TABlock) that jointly leverage spatial-frequency domain analysis and hierarchical degradation information to dynamically activate optimal subnetworks for restoration. Extensive experiments demonstrate that IMDNet significantly outperforms state-of-the-art methods on multi-degradation image restoration, while maintaining competitive performance on single-degradation benchmarks—validating its strong generalization capability and practical utility.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Image restoration (IR) aims to recover clean images from degraded observations. Despite remarkable progress, most existing methods focus on a single degradation type, whereas real-world images often suffer from multiple coexisting degradations, such as rain, noise, and haze coexisting in a single image, which limits their practical effectiveness. In this paper, we propose an adaptive multi-degradation image restoration network that reconstructs images by leveraging decoupled representations of degradation ingredients to guide path selection. Specifically, we design a degradation ingredient decoupling block (DIDBlock) in the encoder to separate degradation ingredients statistically by integrating spatial and frequency domain information, enhancing the recognition of multiple degradation types and making their feature representations independent. In addition, we present fusion block (FBlock) to integrate degradation information across all levels using learnable matrices. In the decoder, we further introduce a task adaptation block (TABlock) that dynamically activates or fuses functional branches based on the multi-degradation representation, flexibly selecting optimal restoration paths under diverse degradation conditions. The resulting tightly integrated architecture, termed IMDNet, is extensively validated through experiments, showing superior performance on multi-degradation restoration while maintaining strong competitiveness on single-degradation tasks.

Problem

• Restores images with multiple coexisting degradations like rain, noise, and haze
• Decouples degradation ingredients using spatial and frequency domain information
• Dynamically selects optimal restoration paths based on degradation conditions

Innovation

• Decouples degradation ingredients via spatial and frequency domains
• Integrates degradation information using learnable fusion matrices
• Dynamically selects restoration paths with task-adaptive blocks