COMMA: Coordinate-aware Modulated Mamba Network for 3D Dispersed Vessel Segmentation

2025-03-04

AI Summary

To address spatial localization ambiguity and insufficient long-range dependency modeling in 3D vascular segmentation, this work proposes a dual-branch network. The global branch employs a channel-compressed Mamba (ccMamba) to capture volumetric long-range contextual dependencies, while the local branch introduces a coordinate-aware modulation (CaM) module to explicitly integrate spatial priors for improved small-vessel localization. This is the first study to incorporate state space models into 3D vascular segmentation. We also construct the largest publicly available 3D vascular dataset to date, comprising 570 cases. Evaluated across six diverse datasets—spanning two imaging modalities and five vascular tissue types—our method achieves state-of-the-art performance: +3.2% Dice score for small vessels, 2.1× faster inference speed than 3D U-Net, and 37% reduced GPU memory consumption.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: Localization, Mapping, and Navigation

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Accurate segmentation of 3D vascular structures is essential for various medical imaging applications. The dispersed nature of vascular structures leads to inherent spatial uncertainty and necessitates location awareness, yet most current 3D medical segmentation models rely on the patch-wise training strategy that usually loses this spatial context. In this study, we introduce the Coordinate-aware Modulated Mamba Network (COMMA) and contribute a manually labeled dataset of 570 cases, the largest publicly available 3D vessel dataset to date. COMMA leverages both entire and cropped patch data through global and local branches, ensuring robust and efficient spatial location awareness. Specifically, COMMA employs a channel-compressed Mamba (ccMamba) block to encode entire image data, capturing long-range dependencies while optimizing computational costs. Additionally, we propose a coordinate-aware modulated (CaM) block to enhance interactions between the global and local branches, allowing the local branch to better perceive spatial information. We evaluate COMMA on six datasets, covering two imaging modalities and five types of vascular tissues. The results demonstrate COMMA's superior performance compared to state-of-the-art methods with computational efficiency, especially in segmenting small vessels. Ablation studies further highlight the importance of our proposed modules and spatial information. The code and data will be open source at https://github.com/shigen-StoneRoot/COMMA.

Problem

• Accurate 3D vascular segmentation for medical imaging applications.
• Addressing spatial uncertainty in dispersed vascular structures.
• Enhancing spatial context awareness in 3D segmentation models.

Innovation

• Coordinate-aware Modulated Mamba Network (COMMA)
• Channel-compressed Mamba (ccMamba) block
• Coordinate-aware modulated (CaM) block