CoDe-SSM: Context-Detail Decoupled State Space Model for Efficient UHD Image Restoration

2026-07-31

AI Summary

Ultra-high-definition (UHD) image restoration faces a fundamental trade-off between computational efficiency and the preservation of high-frequency details. To address this challenge, this work proposes a context-detail decoupled modeling framework that uniquely integrates state space models with an explicit decoupling mechanism. Specifically, a Global Context Scanning Module (GCSM) captures long-range contextual dependencies, while a Local High-Frequency Module (LHFM)—guided by a high-frequency mask and powered by a sparse convolutional mixture-of-experts network—recovers fine-grained details. Evaluated across five UHD benchmarks and five types of degradation tasks, the proposed method consistently outperforms existing approaches, achieving state-of-the-art restoration quality without compromising computational efficiency.

Abstract

Ultra-high-definition (UHD) image restoration must balance the aggregation of spatially recurring degradation cues with the preservation of localized image structures. Compact aggregation can reduce redundant processing but may attenuate edges, textures, and other fine structures. Existing approaches manage UHD restoration cost through downsampling, window partitioning, or cluster-based token reduction; yet many of them do not explicitly retain information that is poorly represented by shared aggregation. In this study, we propose a Context-Detail Decoupled State Space Model (CoDe-SSM) for UHD restoration, which processes aggregated context and clustering residuals in separate pathways. The context modeling pathway, implemented by the Global Cluster Scan Module (GCSM), aggregates features into $K$ input-dependent cluster centers and applies selective SSM reasoning over the resulting fixed-order sequence, enabling cross-region context sharing while decoupling computational cost from spatial resolution. The detail recovery pathway, implemented by the Local High-Frequency Module (LHFM), processes the clustering residual with an input-derived high-frequency mask and a sparse mixture of convolutional experts. Extensive experiments on five UHD benchmarks and five degradation types demonstrate that our explicit context-detail decoupling strategy yields substantial gains in restoration quality while maintaining desirable efficiency.

Problem

• UHD image restoration
• context-detail decoupling
• spatial degradation
• fine structure preservation
• efficient restoration

Innovation

• State Space Model
• Context-Detail Decoupling
• UHD Image Restoration
• Cluster-based Aggregation
• High-Frequency Detail Recovery