HiFi-Mamba: Dual-Stream W-Laplacian Enhanced Mamba for High-Fidelity MRI Reconstruction

2025-08-07

AI Summary

To address the challenge of simultaneously modeling low-frequency structural information and preserving high-frequency details in undersampled k-space MRI reconstruction, this paper proposes a dual-stream Mamba architecture enhanced with a W-Laplacian module. The method introduces: (1) a W-Laplacian module that enables fidelity-driven spectral decoupling, explicitly separating low-frequency semantics from high-frequency textures; (2) a dual-stream Mamba design with an adaptive state-space modulation mechanism to jointly capture long-range dependencies and local fine-grained details; and (3) a unidirectional sequence scanning strategy that maintains modeling capacity while reducing computational redundancy. With linear time complexity, the proposed approach significantly outperforms mainstream CNNs, Transformers, and existing Mamba variants. It achieves state-of-the-art reconstruction accuracy on multiple standard MRI benchmarks using a substantially more lightweight architecture.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Multi-modal Vision
• Cognitive Modeling & Cognitive Systems: (Computational) Cognitive Architectures

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Reconstructing high-fidelity MR images from undersampled k-space data remains a challenging problem in MRI. While Mamba variants for vision tasks offer promising long-range modeling capabilities with linear-time complexity, their direct application to MRI reconstruction inherits two key limitations: (1) insensitivity to high-frequency anatomical details; and (2) reliance on redundant multi-directional scanning. To address these limitations, we introduce High-Fidelity Mamba (HiFi-Mamba), a novel dual-stream Mamba-based architecture comprising stacked W-Laplacian (WL) and HiFi-Mamba blocks. Specifically, the WL block performs fidelity-preserving spectral decoupling, producing complementary low- and high-frequency streams. This separation enables the HiFi-Mamba block to focus on low-frequency structures, enhancing global feature modeling. Concurrently, the HiFi-Mamba block selectively integrates high-frequency features through adaptive state-space modulation, preserving comprehensive spectral details. To eliminate the scanning redundancy, the HiFi-Mamba block adopts a streamlined unidirectional traversal strategy that preserves long-range modeling capability with improved computational efficiency. Extensive experiments on standard MRI reconstruction benchmarks demonstrate that HiFi-Mamba consistently outperforms state-of-the-art CNN-based, Transformer-based, and other Mamba-based models in reconstruction accuracy while maintaining a compact and efficient model design.

Problem

• Reconstructing high-fidelity MR images from undersampled data
• Enhancing sensitivity to high-frequency anatomical details in MRI
• Reducing redundancy in multi-directional scanning for MRI reconstruction

Innovation

• Dual-stream Mamba architecture with W-Laplacian blocks
• Adaptive state-space modulation for high-frequency features
• Streamlined unidirectional traversal for computational efficiency