Flatten Wisely: How Patch Order Shapes Mamba-Powered Vision for MRI Segmentation

2025-07-15

AI Summary

This study reveals the significant impact of patch scan order—a frequently overlooked design choice—on Vision Mamba’s performance in medical MRI segmentation. To address path dependency arising from 2D-to-1D image serialization, we propose the parameter-free, computation-free Multi-Scan 2D (MS2D) module, enabling flexible switching among 21 distinct scanning strategies. We provide the first systematic empirical validation of scan order as a critical implicit hyperparameter: Friedman tests across three public MRI datasets demonstrate performance variations of up to 27 Dice points across scanning paths, with spatially contiguous raster scanning consistently yielding optimal results. Our findings establish a reproducible, zero-cost optimization paradigm and evidence-based design guidelines for Mamba-style vision models in medical imaging.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• User Modeling, Personalization and Recommendation: Practical large-scale studies of user experience
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Vision Mamba models promise transformer-level performance at linear computational cost, but their reliance on serializing 2D images into 1D sequences introduces a critical, yet overlooked, design choice: the patch scan order. In medical imaging, where modalities like brain MRI contain strong anatomical priors, this choice is non-trivial. This paper presents the first systematic study of how scan order impacts MRI segmentation. We introduce Multi-Scan 2D (MS2D), a parameter-free module for Mamba-based architectures that facilitates exploring diverse scan paths without additional computational cost. We conduct a large-scale benchmark of 21 scan strategies on three public datasets (BraTS 2020, ISLES 2022, LGG), covering over 70,000 slices. Our analysis shows conclusively that scan order is a statistically significant factor (Friedman test: $χ^{2}_{20}=43.9, p=0.0016$), with performance varying by as much as 27 Dice points. Spatially contiguous paths -- simple horizontal and vertical rasters -- consistently outperform disjointed diagonal scans. We conclude that scan order is a powerful, cost-free hyperparameter, and provide an evidence-based shortlist of optimal paths to maximize the performance of Mamba models in medical imaging.

Problem

• Impact of patch scan order on MRI segmentation accuracy
• Optimal scan path selection for Mamba-based vision models
• Performance variation due to patch serialization in medical imaging

Innovation

• Introduces MS2D for Mamba-based MRI segmentation
• Systematically studies patch scan order impact
• Identifies optimal scan paths for performance