Progressive Split Mamba: Effective State Space Modelling for Image Restoration

2026-03-10

AI Summary

This work addresses the limitations of existing state space models in image restoration, which suffer from spatial topology disruption due to sequential processing and struggle to balance local detail preservation with long-range consistency. To overcome these issues, we propose a topology-aware hierarchical state space framework that constructs multi-scale representations through a geometrically consistent progressive partitioning strategy—ranging from binary to octal splits—and incorporates cross-scale low-frequency direct connections to retain global context. The method effectively mitigates topological distortion and long-range information decay while maintaining linear computational complexity. Extensive experiments on super-resolution, denoising, and JPEG artifact removal demonstrate consistent and significant performance gains over current Mamba-based and attention-based models.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Learning with Manifolds
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Image restoration requires simultaneously preserving fine-grained local structures and maintaining long-range spatial coherence. While convolutional networks struggle with limited receptive fields, and Transformers incur quadratic complexity for global attention, recent State Space Models (SSMs), such as Mamba, provide an appealing linear-time alternative for long-range dependency modelling. However, naively extending Mamba to 2D images exposes two intrinsic shortcomings. First, flattening 2D feature maps into 1D sequences disrupts spatial topology, leading to locality distortion that hampers precise structural recovery. Second, the stability-driven recurrent dynamics of SSMs induce long-range decay, progressively attenuating information across distant spatial positions and weakening global consistency. Together, these effects limit the effectiveness of state-space modelling in high-fidelity restoration. We propose Progressive Split-Mamba (PS-Mamba), a topology-aware hierarchical state-space framework designed to reconcile locality preservation with efficient global propagation. Instead of sequentially flattening entire feature maps, PS-Mamba performs geometry-consistent partitioning, maintaining neighbourhood integrity prior to state-space processing. A progressive split hierarchy (halves, quadrants, octants) enables structured multi-scale modelling while retaining linear complexity. To counteract long-range decay, we introduce symmetric cross-scale shortcut pathways that directly transmit low-frequency global context across hierarchical levels, stabilising information flow over large spatial extents. Extensive experiments on super-resolution, denoising, and JPEG artifact reduction show consistent improvements over recent Mamba-based and attention-based models with a clear margin.

Problem

• Image Restoration
• State Space Models
• Spatial Coherence
• Local Structure Preservation
• Long-range Dependency

Innovation

• State Space Models
• Image Restoration
• Progressive Split-Mamba
• Long-range Dependency
• Topology-aware Partitioning