Retinal Vessel Segmentation via Morphological Refinement and Adaptive Late Fusion
DOI:
https://doi.org/10.62411/jcta.16093Keywords:
Adaptive late fusion, Fundus imaging, Image segmentation, Local adaptive thresholding, Morphological refinement, Retinal vessel segmentation, Training-free segmentation, Unsupervised segmentationAbstract
Retinal vessel segmentation is fundamental for quantitative retinal vascular analysis; however, accurate delineation remains challenging because of nonuniform illumination, low-contrast capillaries, pathological lesions, and variations in image acquisition. This study presents a rule-based unsupervised retinal vessel segmentation framework based on decision-level adaptive late fusion, in which six complementary local adaptive thresholding methods are integrated and subsequently refined through luminance validation, hysteresis reconstruction, component cleanup, elongation filtering, and boundary smoothing. In this context, unsupervised denotes that no statistical segmentation model is trained using manual vessel annotations; instead, fixed parameters and fusion weights are determined from a small development subset, while all evaluation images remain unseen during method development. The proposed framework was evaluated on 135 independent retinal fundus images from the DRIVE, STARE, CHASE_DB1, HRF, and LES-AV datasets using an eroded field-of-view protocol. It achieved an image-weighted Dice coefficient of 0.7104 (95% bootstrap confidence interval: 0.7005–0.7205), an IoU of 0.5539, a sensitivity of 0.7544, a specificity of 0.9597, a balanced accuracy of 0.8570, and a clDice score of 0.7395. Compared with fixed majority voting, the proposed adaptive late fusion strategy significantly improved segmentation performance in 128 of 135 test images, yielding a mean Dice improvement of 0.0270 (paired Wilcoxon, Holm-adjusted p = 7.67 × 10⁻²¹). Although segmentation performance remains limited for complex pathological images, particularly in the HRF dataset, the proposed framework demonstrates consistent cross-dataset performance, transparent decision-making, and strong reproducibility, providing an effective alternative when interpretable, training-free retinal vessel segmentation is required.References
I. Inamullah, I. Razzak, and S. Jameel, “The eye as a window to systemic health: A survey of retinal imaging from classical techniques to oculomics,” J. Precis. Med. Heal. Dis., vol. 4, p. 100023, Dec. 2025, doi: 10.1016/j.premed.2025.100023.
M. N. Getahun, O. Y. Rogov, D. V. Dylov, A. Somov, A. Bouridane, and R. Hamoudi, “FS-Net: Full scale network and adaptive threshold for improving extraction of micro-retinal vessel structures,” Pattern Recognit. Lett., vol. 189, pp. 188–194, Mar. 2025, doi: 10.1016/j.patrec.2025.01.019.
N. Popovic, M. Radunovic, J. Badnjar, and T. Popovic, “Fractal dimension and lacunarity analysis of retinal microvascular morphology in hypertension and diabetes,” Microvasc. Res., vol. 118, pp. 36–43, Jul. 2018, doi: 10.1016/j.mvr.2018.02.006.
F. Yan, Y. Liu, Q. Zhao, and G. He, “A multi-task deep learning pipeline integrating vessel segmentation and radiomics for multiclass retinal disease classification,” Photodiagnosis Photodyn. Ther., vol. 56, p. 105209, Dec. 2025, doi: 10.1016/j.pdpdt.2025.105209.
Z. Ma and X. Li, “An improved supervised and attention mechanism-based U-Net algorithm for retinal vessel segmentation,” Comput. Biol. Med., vol. 168, p. 107770, Jan. 2024, doi: 10.1016/j.compbiomed.2023.107770.
Y. Wu, C. S. Lee, and A. Y. Lee, “Beyond convolutions and supervised learning with transformers and representation learning for retinal image analysis,” Prog. Retin. Eye Res., vol. 110, p. 101419, Jan. 2026, doi: 10.1016/j.preteyeres.2025.101419.
J. Liu et al., “Unsupervised domain adaptation multi-level adversarial learning-based crossing-domain retinal vessel segmentation,” Comput. Biol. Med., vol. 178, p. 108759, Aug. 2024, doi: 10.1016/j.compbiomed.2024.108759.
W. Ben Messaoud, S. Ben Brahim, A. Zrelli, and M. Yahia, “Domain-specific data augmentation for retinal disease classification using hybrid CNN-Transformer architectures,” Biomed. Signal Process. Control, vol. 121, p. 110325, Aug. 2026, doi: 10.1016/j.bspc.2026.110325.
J. Van Eijgen et al., “Retinal vessel analysis to assess microvascular function in the healthy eye: A systematic review on the response to acute physiological and pathological stressors,” Surv. Ophthalmol., vol. 70, no. 2, pp. 200–214, Mar. 2025, doi: 10.1016/j.survophthal.2024.11.008.
Z. Li, H. Wang, N. Tang, R. Lan, B. Li, and X. Luo, “A Dual-Layer Semantic Fusion Network for Retinal Vessel Segmentation,” Biomed. Signal Process. Control, vol. 112, p. 108463, Feb. 2026, doi: 10.1016/j.bspc.2025.108463.
N. Alsharabi, A. Echtioui, A. Alayba, G. Alshammari, and H. Hamam, “Advanced hybrid UNet architectures for retinal vessel segmentation,” Egypt. Informatics J., vol. 32, p. 100823, Dec. 2025, doi: 10.1016/j.eij.2025.100823.
H. Hu, H. Xiong, F. Long, M. S. Alam, and J. Sang, “Multi-scale Vision Mamba-UNet: A Mamba-based method for retinal vessel segmentation,” Biomed. Signal Process. Control, vol. 112, p. 108435, Feb. 2026, doi: 10.1016/j.bspc.2025.108435.
S. Aslani and H. Sarnel, “A new supervised retinal vessel segmentation method based on robust hybrid features,” Biomed. Signal Process. Control, vol. 30, pp. 1–12, Sep. 2016, doi: 10.1016/j.bspc.2016.05.006.
S. Das, S. Chakraborty, M. Mishra, and S. Majumder, “Assessment of retinal blood vessel segmentation using U-Net model: A deep learning approach,” Franklin Open, vol. 8, p. 100143, Sep. 2024, doi: 10.1016/j.fraope.2024.100143.
L. Ma, J.-X. Jiang, S. Zeng, Y. Li, and Z. Wang, “Frequency view feature interpolation and disentanglement for single-source domain generalization in retinal vessel segmentation,” Biomed. Signal Process. Control, vol. 112, p. 108814, Feb. 2026, doi: 10.1016/j.bspc.2025.108814.
S. Guo, “Enhancing cross-domain generalization in retinal image segmentation via style randomization and style normalization,” Image Vis. Comput., vol. 162, p. 105694, Oct. 2025, doi: 10.1016/j.imavis.2025.105694.
S. Guo and K. Ji, “Random color transformation for single domain generalized retinal image segmentation,” Eng. Appl. Artif. Intell., vol. 136, p. 108907, Oct. 2024, doi: 10.1016/j.engappai.2024.108907.
D. Relan and R. Relan, “Unsupervised sorting of retinal vessels using locally consistent Gaussian mixtures,” Comput. Methods Programs Biomed., vol. 199, p. 105894, Feb. 2021, doi: 10.1016/j.cmpb.2020.105894.
B. Subramani, M. Veluchamy, and A. K. Bhandari, “Optimal color constancy model for biological feature enhancement of retinal fundus images,” Biomed. Signal Process. Control, vol. 117, p. 109637, May 2026, doi: 10.1016/j.bspc.2026.109637.
Q. Xin and L. Han, “Dynamic spatial graph convolution and attention network with dual threshold inference for retinal vessel segmentation,” Biomed. Signal Process. Control, vol. 120, p. 110158, Jul. 2026, doi: 10.1016/j.bspc.2026.110158.
M. Nandy and M. Banerjee, “A Comparative Analysis of Application of Niblack and Sauvola Binarization to Retinal Vessel Segmentation,” in 2017 3rd International Conference on Computational Intelligence and Networks (CINE), Oct. 2017, pp. 105–109. doi: 10.1109/CINE.2017.19.
R. K. Hong et al., “Quantifying the repeatability and reproducibility of Dr. Noon CVD, AI software as medical device for cardiovascular risk assessment via retinal imaging,” Can. J. Ophthalmol., vol. 61, no. 1, pp. 1–11, Feb. 2026, doi: 10.1016/j.jcjo.2025.07.008.
B. C. Fry et al., “Predicting the impact of retinal vessel density on retinal vessel and tissue oxygenation using a theoretical model,” Math. Biosci., vol. 377, p. 109292, Nov. 2024, doi: 10.1016/j.mbs.2024.109292.
X. Shi et al., “The retinal nerve fiber layer mean thickness in patients with early Parkinson’s disease reflects striatal dopamine function,” Neuroscience, vol. 599, pp. 94–102, Apr. 2026, doi: 10.1016/j.neuroscience.2026.02.013.
Z. Chu et al., “Quantification of Choriocapillaris with Phansalkar Local Thresholding: Pitfalls to Avoid,” Am. J. Ophthalmol., vol. 213, pp. 161–176, May 2020, doi: 10.1016/j.ajo.2020.02.003.
D. Benny and K. R. Soumya, “New local adaptive thresholding and dynamic self-organizing feature map techniques for handwritten character recognizer,” in 2015 International Conference on Circuits, Power and Computing Technologies [ICCPCT-2015], Mar. 2015, pp. 1–4. doi: 10.1109/ICCPCT.2015.7159293.
T. Mapayi, S. Viriri, and J.-R. Tapamo, “A New Adaptive Thresholding Technique for Retinal Vessel Segmentation Based on Local Homogeneity Information,” in Lecture Notes in Computer Science, 2014, pp. 558–567. doi: 10.1007/978-3-319-07998-1_64.
L. Du et al., “Deep ensemble learning for accurate retinal vessel segmentation,” Comput. Biol. Med., vol. 158, p. 106829, May 2023, doi: 10.1016/j.compbiomed.2023.106829.
M. Zhang, J. Wang, X. Cao, X. Xu, J. Zhou, and H. Chen, “An integrated global and local thresholding method for segmenting blood vessels in angiography,” Heliyon, vol. 10, no. 22, p. e38579, Nov. 2024, doi: 10.1016/j.heliyon.2024.e38579.
Xiaoyi Jiang and D. Mojon, “Adaptive local thresholding by verification-based multithreshold probing with application to vessel detection in retinal images,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 25, no. 1, pp. 131–137, Jan. 2003, doi: 10.1109/TPAMI.2003.1159954.
C. Zhou, X. Zhang, and H. Chen, “A new robust method for blood vessel segmentation in retinal fundus images based on weighted line detector and hidden Markov model,” Comput. Methods Programs Biomed., vol. 187, p. 105231, Apr. 2020, doi: 10.1016/j.cmpb.2019.105231.
S. Mahapatra, S. Agrawal, P. K. Mishro, and R. B. Pachori, “A novel framework for retinal vessel segmentation using optimal improved frangi filter and adaptive weighted spatial FCM,” Comput. Biol. Med., vol. 147, p. 105770, Aug. 2022, doi: 10.1016/j.compbiomed.2022.105770.
T. M. Khan et al., “Width-wise vessel bifurcation for improved retinal vessel segmentation,” Biomed. Signal Process. Control, vol. 71, p. 103169, Jan. 2022, doi: 10.1016/j.bspc.2021.103169.
Z. Qu, L. Zhuo, J. Cao, X. Li, H. Yin, and Z. Wang, “TP-Net: Two-Path Network for Retinal Vessel Segmentation,” IEEE J. Biomed. Heal. Informatics, vol. 27, no. 4, pp. 1979–1990, Apr. 2023, doi: 10.1109/JBHI.2023.3237704.
D. Hu, H. Li, H. Liu, and I. Oguz, “Domain generalization for retinal vessel segmentation via Hessian-based vector field,” Med. Image Anal., vol. 95, p. 103164, Jul. 2024, doi: 10.1016/j.media.2024.103164.
Z. Tu, Q. Zhou, H. Zou, and X. Zhang, “A Multi-Task Dense Network with Self-Supervised Learning for Retinal Vessel Segmentation,” Electronics, vol. 11, no. 21, p. 3538, Oct. 2022, doi: 10.3390/electronics11213538.
L. Tong et al., “LiViT-Net: A U-Net-like, lightweight Transformer network for retinal vessel segmentation,” Comput. Struct. Biotechnol. J., vol. 24, pp. 213–224, Dec. 2024, doi: 10.1016/j.csbj.2024.03.003.
J. Staal, M. D. Abramoff, M. Niemeijer, M. A. Viergever, and B. van Ginneken, “Ridge-Based Vessel Segmentation in Color Images of the Retina,” IEEE Trans. Med. Imaging, vol. 23, no. 4, pp. 501–509, Apr. 2004, doi: 10.1109/TMI.2004.825627.
A. D. Hoover, V. Kouznetsova, and M. Goldbaum, “Locating blood vessels in retinal images by piecewise threshold probing of a matched filter response,” IEEE Trans. Med. Imaging, vol. 19, no. 3, pp. 203–210, Mar. 2000, doi: 10.1109/42.845178.
M. M. Fraz et al., “An Ensemble Classification-Based Approach Applied to Retinal Blood Vessel Segmentation,” IEEE Trans. Biomed. Eng., vol. 59, no. 9, pp. 2538–2548, Sep. 2012, doi: 10.1109/TBME.2012.2205687.
J. Odstrcilik et al., “Retinal vessel segmentation by improved matched filtering: evaluation on a new high‐resolution fundus image database,” IET Image Process., vol. 7, no. 4, pp. 373–383, Jun. 2013, doi: 10.1049/iet-ipr.2012.0455.
J. I. Orlando, J. Barbosa Breda, K. van Keer, M. B. Blaschko, P. J. Blanco, and C. A. Bulant, “Towards a Glaucoma Risk Index Based on Simulated Hemodynamics from Fundus Images,” in Lecture Notes in Computer Science, 2018, pp. 65–73. doi: 10.1007/978-3-030-00934-2_8.
S. Shit et al., “clDice - a Novel Topology-Preserving Loss Function for Tubular Structure Segmentation,” in 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2021, pp. 16555–16564. doi: 10.1109/CVPR46437.2021.01629.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Afrig Aminuddin, Mohammad Badrul Alam Miah, Ahmed Adil Nafea, Hesmeralda Rojas Enriquez
This work is licensed under a Creative Commons Attribution 4.0 International License.
