OmniSEG: A Multi-Model Ensemble Framework for Retinal Layer Segmentation in OCT Images

Spatial Normalization for Cross-Domain Retinal Layer Segmentation in Optical Coherence Tomography

Retinal layer segmentation in Optical Coherence Tomography (OCT) is a fundamental step for extracting quantitative biomarkers of retinal structure. Indeed, there is a growing interest in the analysis of OCTs in the context of neurodegenerative diseases. However, segmentation remains challenging due to speckle noise, shadowing artifacts, low contrast between adjacent layers, anatomical variability across subjects, and domain shifts arising from different acquisition protocols and clinical populations. While deep learning methods have achieved remarkable performance, their robustness and generalization across heterogeneous datasets remain limited. In this work, we investigate the role of spatial normalization as a preprocessing strategy to mitigate geometric domain shifts and improve the consistency of retinal layer segmentation. Inspired by standard practices in neuroimaging, we introduce a fovea-centered normalization framework that aligns OCT volumes into a common anatomical reference. We perform a comprehensive evaluation of state-of-the-art deep learning architectures. To provide a comprehensive assessment of segmentation quality, we combine conventional overlap-based metrics at B-scan level with topology-aware metrics at A-scan level and thickness-based measures at the en-face level. In cases where a ground truth is not available, we propose topology violation quantitative metrics that do not require ground truth annotations and a thickness-based qualitative assessment that captures structural consistency and clinically relevant patterns at the en-face level. The results demonstrate the importance of spatial normalization in OCT segmentation pipelines toward the development of robust and clinically meaningful retinal analysis tools, enabling reliable biomarker extraction and downstream computational analysis in neurodegenerative research.

Domain-Adaptive Retinal Vessel Segmentation for Unannotated Fundus Images

Accurate retinal vessel segmentation supports quantitative vascular analysis in assessing ocular and systemic diseases. Yet, its clinical scalability is constrained by limited pixel-level annotations and domain shift across heterogeneous fundus datasets. Thus, this study proposes a domain-adaptive retinal vessel segmentation model (DA-VesselNet), a weakly supervised approach that transfers vessel-segmentation knowledge from annotated source datasets to the unannotated Retinal Fundus Multi-Disease Image Dataset (RFMiD). The model was trained on DRIVE, CHASE_DB1, and FIVES, and adapted to RFMiD. The ResNet50 encoder with an attention-gated U-Net decoder, confidence-aware pseudo-label supervision, and perturbation consistency regularisation were used for the training process. Results on the held-out CHASE_DB1 indicated that DA-VesselNet achieved a Dice score of 0.5778, an Intersection over Union (IoU) of 0.4082, and an Area Under the Curve (AUC) of 0.9518. On 200 held-out FIVES test images with different pathological features, it achieved a Dice score of 0.7447 and an AUC of 0.9733, outperforming the source-only baseline model. To assess adaptation independently of source-adjacent data, the model was further tested on STARE and HRF, two domains excluded entirely from source training. Adaptation improved Dice by 0.0469 on STARE and 0.0056 on HRF with AUC gains of 0.0140 and 0.0121, respectively. Ablation analysis identified source-anchored supervision as the dominant contributor to performance. The adapted model was subsequently applied to generate vessel pseudo-labels for the RFMiD target domain, providing a structural resource for future vessel-informed analysis. These findings demonstrate that DA-VesselNet offers a scalable solution for creating clinically relevant pseudo-labelled vessels in fundus imaging with limited annotations.

Deep Learning-based Automated Segmentation of Ultra-Widefield Retinal Vasculature for Cardiometabolic Disease Association Analysis.

Retinal vessel analysis in ultra-widefield (UWF) images provides a unique opportunity for large-scale assessment of systemic microvascular health. However, accurate segmentation in true-color UWF images remains challenging due to the large field of view, complex background, and reduced vessel contrast. To address these challenges, we develop ECS-Net, a dedicated deep learning (DL) framework for retinal vessel segmentation in true-color UWF images. Our ECS-Net adopts an enhanced encoder decoder architecture that integrates a dual-domain context enhancement module (DCEM), consisting of an anisotropic spatial focus unit (ASFU) and a feature correlation calibration unit (FCCU), together with atrous spatial pyramid pooling (ASPP) for multi-scale context modeling. The proposed ECS Net achieved a higher Dice coefficient of 0.8349 than other state-of-the-art algorithms (0.7001-0.8136), demonstrating strong generalization under real-world imaging conditions. Building upon accurate vessel extraction, region-specific vascular parameters, including vessel density (VD), fractal dimension (FD), tortuosity (TC), and mean curvature (MC), were quantified separately for central (45°), peripheral (45° 133°), and global (133°) zones. Multivariable logistic regression was used to evaluate associations with metabolic diseases in 4,618 participants. Hypertension showed significant inverse associations with VD and FD across all zones (all P < 0.001). Diabetes exhibited striking regional specificity, characterized by increased TC and MC and decreased FD were confined to the peripheral and global zones (all P< 0.01), with inverse associations for TC and MC detectable only in the global zone (P < 0.05). This study establishes the first DL framework specifically designed for retinal vessel segmentation in true color UWF images and reveals disease-specific regional vascular patterns that would be missed by conventional fundus photography, highlighting the value of UWF imaging for comprehensive systemic disease assessment and providing a biological prior for the development of interpretable and region-aware models. The code will be released on GitHub: https://github.com/yxyXinyue/UWF Retinal-Vasculature-Segmentation-Cardiometabolic.

Semantic Segmentation of Optical Coherence Tomography Images Based on TUnet+ Modeling

Accurate segmentation of Optical Coherence Tomography images is critical for assisting clinicians in identifying lesion regions and evaluating disease progression in retinal and cardiovascular disorders. As OCT is an optical electromagnetic-wave imaging modality, segmentation performance is closely related to the interpretation of wave scattering, tissue-layer boundaries, and propagation-induced image features. Existing OCT segmentation models are often limited by high computational complexity and insufficient accuracy, which restrict efficient and reliable clinical application. To address these challenges, this study proposes a lightweight TUnet+ network incorporating an innovatively designed PSE_C2fCIB module. Extensive experiments were conducted on the publicly available GOALS2022 dataset, including 200 deidentified OCT images augmented for training and validation. The results show that TUnet+ achieves state-of-the-art performance, with an accuracy of 99.3%, an F1-score of 95.65%, and a mean Intersection over Union of 91.87%. Through multi-scale feature fusion, channel attention, and efficient contextual modeling, the proposed framework improves the recognition of fine structural boundaries in OCT images. The method provides a robust technical foundation for automated diagnostic support and quantitative disease assessment, and it also demonstrates the value of electromagnetic-wave imaging analysis in biomedical engineering.

Available View Set-guided multi-view retinal disease classification from OCT and OCTA.

Conditional Flow Matching-based Semi-supervised Segmentation for Quality Assessment of Fundus Images.

Axial misalignment in fundus imaging can severely distort anatomical geometry, frequently reducing the reliability of clinical diagnoses. The orientation of the Disc-Fovea Line (DFL) remains the best method for rotation quality control(QC), but it can be challenging to locate the Optic Disc (OD) and macula, which weakens its utility. The weak textures and few pixel-level labels bring great difficulties for macula annotation, resulting in a scarcity of available annotation data. To address this issue, we intro duceFMSSL-GAN, anannotation-efficientSemi-Supervised Semantic Segmentation (SSSS) framework. In particular, a Conditional Flow Matching (CFM) module is added to turn rough, noisy predictions into anatomically correct struc = tures by learning a continuous evolutionary vector field. Our CFM-based approach directly models the probability path, which makesinference much faster while still keeping high-precision structural refinement. This is different from traditional diffusion models, which have slow, multi-step iterative denoising. A Generative Adversarial Network (GAN) is added to impose topological constraints at the distribution level, which enables the spatial relationship between landmarks to remain consistent. For the assessment, our system automatically marks the images as non-compliant when the calculated OD-macula angle is more than 15°. Extensive experiments on a private dataset (i.e., the Szeye) and two public datasets (i.e., REFUGE and ORIGA) demonstrate that the proposed method performs well with a wide range of labeled data ratios and outperforms other compared semi-supervised methods. Our framework provides an annotation-efficient solution for objective fundus image quality assessment, demonstrating a high concordance with expert measurements.