MM-UNet: Morph Mamba U-shaped Convolutional Networks for Retinal Vessel Segmentation

2025-11-04

AI Summary

Retinal vessel segmentation faces accuracy and robustness bottlenecks due to fine-scale structures, complex branching patterns, and substantial inter-image morphological variability. To address these challenges, we propose a novel U-shaped network architecture. Our key contributions are: (1) Morph Mamba convolutional layers that enhance topological awareness of vascular structures; (2) a reverse selective state-guidance module that improves geometric boundary localization accuracy and decoding efficiency; and (3) synergistic integration of convolutional modeling and state-space sequence modeling to jointly optimize local details and global morphology. Evaluated on the DRIVE and STARE benchmarks, our method achieves absolute F1-score improvements of 1.64% and 1.25%, respectively, surpassing state-of-the-art approaches. The implementation is publicly available.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Multimodal Learning
• Intelligent Robots: State Estimation

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Accurate detection of retinal vessels plays a critical role in reflecting a wide range of health status indicators in the clinical diagnosis of ocular diseases. Recently, advances in deep learning have led to a surge in retinal vessel segmentation methods, which have significantly contributed to the quantitative analysis of vascular morphology. However, retinal vasculature differs significantly from conventional segmentation targets in that it consists of extremely thin and branching structures, whose global morphology varies greatly across images. These characteristics continue to pose challenges to segmentation precision and robustness. To address these issues, we propose MM-UNet, a novel architecture tailored for efficient retinal vessel segmentation. The model incorporates Morph Mamba Convolution layers, which replace pointwise convolutions to enhance branching topological perception through morph, state-aware feature sampling. Additionally, Reverse Selective State Guidance modules integrate reverse guidance theory with state-space modeling to improve geometric boundary awareness and decoding efficiency. Extensive experiments conducted on two public retinal vessel segmentation datasets demonstrate the superior performance of the proposed method in segmentation accuracy. Compared to the existing approaches, MM-UNet achieves F1-score gains of 1.64 % on DRIVE and 1.25 % on STARE, demonstrating its effectiveness and advancement. The project code is public via https://github.com/liujiawen-jpg/MM-UNet.

Problem

• Segmentation of extremely thin and branching retinal vessel structures
• Improving segmentation precision and robustness across varying morphologies
• Enhancing global topology perception and geometric boundary awareness

Innovation

• Morph Mamba Convolution layers enhance branching topological perception
• Reverse Selective State Guidance modules improve geometric boundary awareness
• U-shaped architecture tailored for efficient retinal vessel segmentation