EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Hierarchical Multi-Scale Mamba with Tubular Structure-Aware Convolution for Retinal Vessel Segmentation

Tao Wang, Dongyuan Tian, Haonan Zhao iD, Jiamin Liu, Weijie Wang iD, Chunpei Li, Guixia Liu

Entropy · 14.08.2025 · 27 (8) · 862

DOI: 10.3390/e27080862 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Entropy, online erschienen am 14.08.2025. Verfasst von Tao Wang und 6 weiteren Autor:innen. Beteiligte Einrichtung: College of Computer Science and Technology, Jilin University, Changchun 130012, China. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 62 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Nature Science Foundation of China;, Science and Technology Development Program of Jilin Province, Natural Science Foundation of Xiamen, China.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Retinal vessel segmentation plays a crucial role in diagnosing various retinal and cardiovascular diseases and serves as a foundation for computer-aided diagnostic systems. Blood vessels in color retinal fundus images, captured using fundus cameras, are often affected by illumination variations and noise, making it difficult to preserve vascular integrity and posing a significant challenge for vessel segmentation. In this paper, we propose HM-Mamba, a novel hierarchical multi-scale Mamba-based architecture that incorporates tubular structure-aware convolution to extract both local and global vascular features for retinal vessel segmentation. First, we introduce a tubular structure-aware convolution to reinforce vessel continuity and integrity. Building on this, we design a multi-scale fusion module that aggregates features across varying receptive fields, enhancing the model’s robustness in representing both primary trunks and fine branches. Second, we integrate multi-branch Fourier transform with the dynamic state modeling capability of Mamba to capture both long-range dependencies and multi-frequency information. This design enables robust feature representation and adaptive fusion, thereby enhancing the network’s ability to model complex spatial patterns. Furthermore, we propose a hierarchical multi-scale interactive Mamba block that integrates multi-level encoder features through gated Mamba-based global context modeling and residual connections, enabling effective multi-scale semantic fusion and reducing detail loss during downsampling. Extensive evaluations on five widely used benchmark datasets—DRIVE, CHASE_DB1, STARE, IOSTAR, and LES-AV—demonstrate the superior performance of HM-Mamba, yielding Dice coefficients of 0.8327, 0.8197, 0.8239, 0.8307, and 0.8426, respectively.

Schlagwörter

Stichwörter der Autor:innen

Mambaattention mechanismmulti-scale fusionretinal vessel segmentationtubular structure-aware convolution

Autor:innen und Institutionen

  • College of Computer Science and Technology, Jilin University, Changchun 130012, China.
  • Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun 130012, China.
  • Department of Information Engineering and Computer Science, University of Trento, 38123 Trento, Italy.
  • School of Science, Jimei University, Xiamen 361021, China.
  • School of Computer Science and Engineering, Guangxi Normal University, Guilin 541001, China.

ORCID-Profile: Haonan Zhao, Weijie Wang

Förderung

  • National Nature Science Foundation of China;
  • Science and Technology Development Program of Jilin Province
  • Natural Science Foundation of Xiamen, China
  • Foundation of Fujian Educational Committee
  • National Nature Science Foundation of China
  • Key Laboratory for Symbol Computation and Knowledge Engineering of the National Education Ministry of China, Jilin University

Zitieren

APA

Wang, T., Tian, D., Zhao, H., Liu, J., Wang, W., Li, C., & Liu, G. (2025). Hierarchical Multi-Scale Mamba with Tubular Structure-Aware Convolution for Retinal Vessel Segmentation. Entropy, 27(8), 862. https://doi.org/10.3390/e27080862

Vancouver

Wang T, Tian D, Zhao H, Liu J, Wang W, Li C, et al. Hierarchical Multi-Scale Mamba with Tubular Structure-Aware Convolution for Retinal Vessel Segmentation. Entropy. 2025;27(8):862. doi:10.3390/e27080862

BibTeX
@article{Wang2025,
  author = {Wang, Tao and Tian, Dongyuan and Zhao, Haonan and Liu, Jiamin and Wang, Weijie and Li, Chunpei and Liu, Guixia},
  title = {Hierarchical Multi-Scale Mamba with Tubular Structure-Aware Convolution for Retinal Vessel Segmentation},
  journal = {Entropy},
  year = {2025},
  volume = {27},
  number = {8},
  pages = {862},
  doi = {10.3390/e27080862}
}
Kostenlos starten →