Balancing Efficiency and Restoration: Lightweight Mamba-Based Model for CT Metal Artifact Reduction

2026-04-07

AI Summary

This work addresses the severe artifacts in CT images caused by metallic implants, which lead to structural distortions and pose a persistent challenge for existing methods that struggle to balance reconstruction fidelity with computational efficiency. The authors propose MARMamba, a novel approach that requires only artifact-contaminated CT images as input—eliminating the need for additional data such as sinograms. Built upon a lightweight UNet architecture, MARMamba introduces multi-scale Mamba modules and flip-augmented Mamba blocks to capture long-range contextual information from multiple directions, alongside a max-average feedforward network designed to effectively fuse salient and global features. Experimental results demonstrate that MARMamba significantly outperforms current state-of-the-art methods across multiple quantitative metrics, achieving an optimal trade-off among model parameters, memory consumption, and computational cost while preserving anatomical structures, thereby showing strong potential for clinical deployment.

Abstract

In computed tomography imaging, metal implants frequently generate severe artifacts that compromise image quality and hinder diagnostic accuracy. There are three main challenges in the existing methods: the deterioration of organ and tissue structures, dependence on sinogram data, and an imbalance between resource use and restoration efficiency. Addressing these issues, we introduce MARMamba, which effectively eliminates artifacts caused by metals of different sizes while maintaining the integrity of the original anatomical structures of the image. Furthermore, this model only focuses on CT images affected by metal artifacts, thus negating the requirement for additional input data. The model is a streamlined UNet architecture, which incorporates multi-scale Mamba (MS-Mamba) as its core module. Within MS-Mamba, a flip mamba block captures comprehensive contextual information by analyzing images from multiple orientations. Subsequently, the average maximum feed-forward network integrates critical features with average features to suppress the artifacts. This combination allows MARMamba to reduce artifacts efficiently. The experimental results demonstrate that our model excels in reducing metal artifacts, offering distinct advantages over other models. It also strikes an optimal balance between computational demands, memory usage, and the number of parameters, highlighting its practical utility in the real world. The code of the presented model is available at: https://github.com/RICKand-MORTY/MARMamba.

Problem

• metal artifact reduction
• computed tomography
• image restoration
• structural integrity
• resource efficiency

Innovation

• Mamba
• metal artifact reduction
• lightweight model
• multi-scale context
• CT image restoration