Optical coherence tomography angiography (OCTA) provides non-invasive visualization of retinal microvasculature, but learning robust representations remains challenging due to sparse vessel structures and strong topological constraints. Many existing self-supervised learning approaches, including masked autoencoders, are primarily designed for dense natural images and rely on uniform masking and pixel-level reconstruction, which may inadequately capture vascular geometry. We propose VAMAE, a vessel-aware masked autoencoding framework for self-supervised pretraining on OCTA images. The approach incorporates anatomically informed masking that emphasizes vessel-rich regions using vesselness and skeleton-based cues, encouraging the model to focus on vascular connectivity and branching patterns. In addition, the pretraining objective includes reconstructing multiple complementary targets, enabling the model to capture appearance, structural, and topological information. We evaluate the proposed pretraining strategy on the OCTA-500 benchmark for several vessel segmentation tasks under varying levels of supervision. The results indicate that vessel-aware masking and multi-target reconstruction provide consistent improvements over standard masked autoencoding baselines, particularly in limited-label settings, suggesting the potential of geometry-aware self-supervised learning for OCTA analysis.
Figures & tables
Figure 1: Vessel-aware masking strategy. Input images are processed to extract vesselness maps, binary vessel maps, and vessel skeletons. Patch-wise structural priors (vessel density and skeletal presence) guide adaptive masking, prioritizing vessel-rich regions for self-supervised pre-training.
Figure 2: VAMAE architecture with vessel-aware masking and multi-target reconstruction of intensity, vesselness, and skeleton.
Method
Type
50% Labels
100% Labels
p -value †
Cohen’s d
Random MAE [ 6 ]
General SSL
68.5 ± 1.3
76.8 ± 0.9
baseline
–
MedMAE [ 13 ]
Medical SSL
69.2 ± 1.1
77.2 ± 0.8
0.113
0.05
BioVessel-Net [ 4 ]
OCTA SSL
72.1 ± 1.0
78.9 ± 0.7
0.002*
0.29
Pissas et al. [ 14 ]
Retinal SSL
73.8 ± 0.9
79.7 ± 0.6
0.001*
0.38
VAMAE (ours)
OCTA SSL
78.4 ± 0.8
82.4 ± 0.5
0.002*
0.71
† Paired t-test vs. Random MAE at 100% labels. *Significant after Bonferroni correction.
Table 1: Comparison with SSL methods on large vessel segmentation (Dice %).
Figure 3: Qualitative segmentation results comparing VAMAE with MAE, and other baselines. VAMAE better preserves thin vessels and complex bifurcations.
Method
50% Labels
100% Labels
Params
Training
U-Net (from scratch) [ 22 ]
56.8 ± 1.4
69.2 ± 0.9
31M
Supervised
CS 2 -Net [ 23 ]
74.3 ± 1.1
81.7 ± 0.6
45M
Supervised
VAMAE (ours)
78.4 ± 0.8
82.4 ± 0.5
86M
SSL + FT
Table 2: Comparison with supervised methods (Dice %).
Task
50% Labels
75% Labels
100% Labels
p -value †
Large Vessel
78.4 ± 0.8 (+9.9)
80.6 ± 0.6 (+6.3)
82.4 ± 0.5 (+5.6)
<0.001*
FAZ
92.7 ± 0.5 (+4.2)
93.6 ± 0.4 (+3.1)
94.1 ± 0.3 (+2.8)
<0.001*
Vein
61.2 ± 0.9 (+6.1)
64.8 ± 0.7 (+5.7)
67.3 ± 0.6 (+5.1)
<0.001*
† Paired t-test vs. Random MAE at 100% labels.
Table 3: Cross-task generalization on OCTA-500 (Dice %). Gains over Random MAE in parentheses.
Configuration
Dice
Δ vs. Baseline
p -value
Random masking (baseline)
76.8 ± 0.9
0.0
–
+ Density masking
79.2 ± 0.7
+2.4
0.002*
+ Hybrid masking ( α=0.6 )
81.6 ± 0.6
+4.8
<0.001*
Hybrid + Multi-target (I+V+S)
82.4 ± 0.5
+5.6
<0.001*
- Skeleton target
80.7 ± 0.6
+3.9
<0.001*
- Vesselness target
79.2 ± 0.8
+2.4
0.004*
Table 4: Component ablation on large vessels at 100% labels (Dice %).
α
Interpretation
Dice
Δ vs. Random
0.0
Pure random (baseline)
76.8 ± 0.9
0.0
0.4
Moderate vessel bias
80.3 ± 0.7
+3.5*
0.6
Balanced (ours)
82.4 ± 0.5
+5.6*
0.8
Strong vessel bias
81.7 ± 0.6
+4.9*
1.0
Deterministic vessel-only
80.1 ± 0.7
+3.3*
*Significant at p<0.001 vs. α=0.0 .
Table 5: Effect of vessel prioritization weight α (Dice %).
Category
n
Random MAE
VAMAE
Δ / p -value
Healthy
80
79.2 ± 0.8
84.4 ± 0.6
+5.2 / <0.001*
Mild DR
48
76.8 ± 1.0
81.6 ± 0.8
+4.8 / <0.001*
Moderate DR
42
74.1 ± 1.3
79.2 ± 1.0
+5.1 / <0.001*
Severe DR
30
68.7 ± 1.9
71.8 ± 1.6
+3.1 / 0.008*
Table 6: Performance by disease severity (Dice %).
Existing deep learning frameworks for Optical Coherence Tomography Angiography (OCTA) vessel segmentation are largely derived from the U-Net architecture, which serves as the foundation for most current designs. However, most of these methods focus only on holistic representation, struggling to address the problem of low local contrast unique to OCTA, which leads to vessel discontinuities and loss of detail. To address these problems, we propose LSENet, which builds upon the U-Net architecture by introducing three core innovative modules: To address vessel discontinuities, we introduce the Patch Information Enhance module (PIE), which replaces standard skip connections to execute patch-wise attention. To mitigate detail loss, the Multiscale Feature Fusion module (MFF) is proposed to feed the PIE module rich, multi-scale information by extracting visually interpretable features from both the original input and preceding layers. Finally, the Connectivity Refinement Decoder (CRD) is designed to refine features from all levels and utilize a large kernel in the final convolutional layer to reduce fragmentation. Experiments on three public datasets (OCTA-500, ROSE-1, and ROSSA) demonstrate that our proposed LSENet achieves state-of-the-art performance while requiring fewer parameters.
Tuopusen Huang, Ding Ma, Xiangqian Wu
Faculty of Computing Harbin Institute of Technology Harbin, China
Ultra-widefield (UWF) swept-source optical coherence tomography angiography (SS-OCTA) enables large-area retinal vascular imaging, yet vessel segmentation at this scale lacks dedicated public benchmarks and comprehensive evaluation for quantitative vascular analysis. We introduce WOIVES, to our knowledge the first publicly available UWF SS-OCTA vessel-segmentation dataset, comprising 206 eyes from 152 participants with a 24x20mm^2 field of view. WOIVES spans emmetropia to high myopia and provides soft probability vessel annotations. We further propose PG-Mamba, a visual state space model that enhances conventional directional scans with two complementary polar-coordinate scan orders. An auxiliary Dynamic FOV Gating module performs spatial modulation at the bottleneck. PG-Mamba outperformed seven competitive approaches on broad segmentation metrics under cross-validation. It achieved the lowest median absolute errors for vessel density, fractal dimension, and vessel length density. WOIVES is publicly available on Zenodo (DOI: 10.5281/zenodo.21904672), and the PG-Mamba code is available at https://github.com/syb1234567/PG-Mamba.
Yang Liu, Yibing Shen, Keming Zhao +12
Tsinghua Shenzhen International Graduate School, Shenzhen 518109, China · Tsinghua University, Beijing, China · Medical Artificial Intelligence Lab, The First Hospital of Hebei Medical University, Hebei Medical University, Shijiazhuang 050000, China +2
Optical coherence tomographic angiography (OCTA) is a powerful technique for imaging retinal microvasculature. However, acquiring reliable quantification of retinal blood flow and areas of retinal nonperfusion is challenging because of imaging artifacts. Existing methods primarily focus on noise suppression, projection artifact removal, or signal enhancement to improve the image quality of OCTA in cross-sectional or two-dimensional (2D) en face projections, while neglecting the intrinsic three-dimensional vascular architecture. In this study, we propose a deep learning-based algorithm for restoring capillary anatomical vasculature from a single OCTA volume. The network consists of an EfficientNet-B5 encoder and a decoder incorporating concurrent spatial and channel squeeze-and-excitation modules, connected via skip connections to preserve spatial resolution. Three adjacent B-frames are used as input to predict the restored middle B-frame. We evaluated the performance of the model using the peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) against ground truth generated from averaging multiple scans. The results show that the proposed model significantly (both p < 0.001) improved image quality compared with the original single OCTA volume, with a PSNR of 26.16 +/- 1.26 vs. 22.23 +/- 0.78 and an SSIM of 0.91 +/- 0.02 vs. 0.72 +/- 0.03. The proposed model also significantly (p < 0.001) improved microvascular fidelity, measured by the Dice coefficient overlap between the model output and ground truth, in both 2D and 3D by at least 3.8% and 51.2%, respectively, across several different vascular slabs.
Yukun Guo, Min Gao, Tristan T. Hormel +3
Casey Eye Institute, Oregon Health & Science University, Portland, OR 97239, USA · Department of Biomedical Engineering, Oregon Health & Science University, Portland, OR 97239, USA