EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

SDA-SwinNet: Swin-UNet with Dense Skip and Shift-ASPP for Retinal Vessel Segmentation

Jinghua Xiao, Ming Zhao iD, Rui Yang, Zhengqiang Wang, Tie Luo iD, Wangyu Wu iD

Sensors · 03.08.2026 · 26 (15) · 4899

DOI: 10.3390/s26154899 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 03.08.2026. Verfasst von Jinghua Xiao und 5 weiteren Autor:innen. Beteiligte Einrichtung: School of Computer Science, Yangtze University, Jingzhou 434025, China. Hauptthemen laut MEDLINE: Retinal Vessels, Image Processing, Computer-Assisted. Die Arbeit nennt 36 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Wuxi University Research Start-up Fund for High-level Talents.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Retinal artery/vein segmentation is a prerequisite for many ophthalmic diagnostic tools. Yet, the task remains difficult: vessels form complex trees, vary widely in caliber, and often appear low-contrast at terminal branches. We propose SDA-SwinNet to handle these challenges. The network adopts Swin-UNet as its backbone and adds three modifications: a Shift-ASPP module for multi-scale context, an HF-Bridge for cross-level feature fusion, and a fractal-constrained loss with a differentiable topology surrogate. Experimental results on the DRIVE-AV and LES-AV datasets show that the proposed model achieves an overall F1-score of 73.13% on DRIVE-AV and 67.85% on LES-AV, with additional class-wise evaluations for arteries and veins. The results demonstrate that SDA-SwinNet achieves a competitive trade-off between segmentation accuracy and computational efficiency.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Retinal VesselsImage Processing, Computer-AssistedHumansAlgorithmsDeep LearningRetinal Artery

Stichwörter der Autor:innen

ASPPSwin transformer encoderSwin-UNetdeep learningfundus vessel segmentationmedical image analysis

Autor:innen und Institutionen

  • School of Computer Science, Yangtze University, Jingzhou 434025, China.
  • School of Cyber Science and Engineering, Wuxi University, Wuxi 214105, China.
  • Wuxi Key Laboratory of Artificial Intelligence and Security, Wuxi University, Wuxi 214105, China.
  • School of Physics & Electronic Engineering, Hubei University of Arts and Science, Xiangyang 441053, China.
  • School of Computer Science, University of Liverpool, Liverpool L69 3DR, UK.

ORCID-Profile: Ming Zhao, Tie Luo, Wangyu Wu

Förderung

  • Wuxi University Research Start-up Fund for High-level Talents

Zitieren

APA

Xiao, J., Zhao, M., Yang, R., Wang, Z., Luo, T., & Wu, W. (2026). SDA-SwinNet: Swin-UNet with Dense Skip and Shift-ASPP for Retinal Vessel Segmentation. Sensors, 26(15), 4899. https://doi.org/10.3390/s26154899

Vancouver

Xiao J, Zhao M, Yang R, Wang Z, Luo T, Wu W. SDA-SwinNet: Swin-UNet with Dense Skip and Shift-ASPP for Retinal Vessel Segmentation. Sensors. 2026;26(15):4899. doi:10.3390/s26154899

BibTeX
@article{Xiao2026,
  author = {Xiao, Jinghua and Zhao, Ming and Yang, Rui and Wang, Zhengqiang and Luo, Tie and Wu, Wangyu},
  title = {SDA-SwinNet: Swin-UNet with Dense Skip and Shift-ASPP for Retinal Vessel Segmentation},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {15},
  pages = {4899},
  doi = {10.3390/s26154899}
}
Kostenlos starten →