Gradient based Severity Labeling for Biomarker Classification in OCT

Kiran Kokilepersaud; Mohit Prabhushankar; Ghassan AlRegib; Stephanie Trejo Corona; Charles Wykoff

cs.CVFeb 23, 2026

Gradient based Severity Labeling for Biomarker Classification in OCT

Authors: Kiran Kokilepersaud, Mohit Prabhushankar, Ghassan AlRegib, Stephanie Trejo Corona, Charles Wykoff

Organizations: OLIVES at the Center for Signal and Information Processing CSIP, School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, USA · Retina Consultants Texas, Retina Consultants of America, Houston, Texas, USA

Abstract

In this paper, we propose a novel selection strategy for contrastive learning for medical images. On natural images, contrastive learning uses augmentations to select positive and negative pairs for the contrastive loss. However, in the medical domain, arbitrary augmentations have the potential to distort small localized regions that contain the biomarkers we are interested in detecting. A more intuitive approach is to select samples with similar disease severity characteristics, since these samples are more likely to have similar structures related to the progression of a disease. To enable this, we introduce a method that generates disease severity labels for unlabeled OCT scans on the basis of gradient responses from an anomaly detection algorithm. These labels are used to train a supervised contrastive learning setup to improve biomarker classification accuracy by as much as 6% above self-supervised baselines for key indicators of Diabetic Retinopathy.

Figures & tables

Figure 1: OCT Scans of biomarkers from the Prime + TREX DME datasets. The biomarkers are A) Intraretinal Hyperreflective Foci (IRHRF), B) Intraretinal Fluid (IRF) and Diabetic Macular Edema (DME) C) Partially Attached Vitreous Face (PAVF), and D) Fully Attached Vitreous Face (FAVF). The white arrows point out examples of where these biomarkers are found in the images. A discussion of biomarkers in OCT can be found at [ 19 ] .

Figure 2: From a healthy manifold learned from a trained auto-encoder, we can compute distance to the manifold of more severely diseased cases via a Severity Score (SS). Severity Score is calculated via some model response and increases as a sample is more anomalous compared to the learned healthy manifold.

Panel of Figure 3 — Figure 3: Overview of severity labeling methodology.

Figure 4: Overview of supervised contrastive learning and linear fine-tuning steps. 1) Supervised Contrastive Loss on generated severity labels from previously unlabeled data. 2) Attach linear layer and train on labeled biomarker data.

Severity Label Training Results (Accuracy / F1-Score)
Method	IRF	DME	IRHRF	FAVF	PAVF	Multi-Label
SimCLR [ 2 ]	75.13% / .715	80.61% / .772	59.03% /.675	75.43% / .761	52.69% / .249	.754
PCL [ 16 ]	76.50 % / .717	80.11% / .761	59.1% /.683	76.30 % / .773	51.40% / .165	.767
Moco v2 [ 3 ]	76.00% / .720	82.24% / .793	59.6% / .692	75.00% / .784	52.5% / .201	.769
$SL_{5000}$	75.20% / .698	81.46% / .786	66.83 % / .695	75.39% / .756	54.7% / .314	.774
$SL_{10000}$	75.46% / .732	82.35% / .796	62.00% / .700	73.56% / .722	56.69 % / .370	.771

Table 1: Performance of severity labeling and supervised contrastive learning approach on individual biomarkers. Multi-Label is the average AUC from the multi-label classification task.

SL_{N}

refers to training an encoder after having divided the training set into

N

severity label bins.

Anomaly Detector Ablation Study
Method	Multi-Label
MSP [ 8 ]	.764
ODIN [ 18 ]	.761
Mahalabonis [ 15 ]	.772
SL	.774

Table 2: Analysis of using different anomaly detectors to generate severity labels. For each method scores were generated for each unlabeled image in the Biomarker dataset and discretized into 5000 bins.

Figure 5: Visual representation of OCT scans with high and low severity scores.

Explore similar work

TRIAGE: Risk-Controlled Pseudo-Label Admission for Annotation-Efficient Semi-Supervised Retinal OCT Classification
Aug 14, 2026Md Ashraful Hossen Akash, Shyla Afroge, Abdullah Al Mamun +2Medical Image Classification Pseudo-Labeling
Department of Computer Science & Engineering, Rajshahi University of Engineering & Technology, Rajshahi-6204, Bangladesh · Elite Research Lab LLC, New York, USA · Faculty of Information Science and Technology, Multimedia University, Melaka, Malaysia +1
Chronological Contrastive Learning: Few-Shot Progression Assessment in Irreversible Diseases
Mar 23, 2026Clemens Watzenböck, Daniel Aletaha, Michaël Deman +9Medical Image Analysis Medical Image Representation Learning
Computational Imaging Research Lab, Department of Biomedical Imaging and Image-guided Therapy, Medical University of Vienna, Vienna, Austria · Comprehensive Center for Artificial Intelligence in Medicine, Medical University of Vienna, Vienna, Austria · Division of Rheumatology, Department of Medicine III, Medical University of Vienna, Vienna, Austria +3
Pixel-level Counterfactual Contrastive Learning for Medical Image Segmentation
Mar 17, 2026Marceau Lafargue-Hauret, Raghav Mehta, Fabio De Sousa Ribeiro +2Medical Image Segmentation Contrastive Learning
Department of Computing, Imperial College London, UK

Aug 14, 2026J/K move · Enter open · S save

TRIAGE: Risk-Controlled Pseudo-Label Admission for Annotation-Efficient Semi-Supervised Retinal OCT Classification

Md Ashraful Hossen Akash, Shyla Afroge, Abdullah Al Mamun, Md. Kishor Morol+1

Department of Computer Science & Engineering, Rajshahi University of Engineering & Technology, Rajshahi-6204, Bangladesh · Elite Research Lab LLC, New York, USA · Faculty of Information Science and Technology, Multimedia University, Melaka, Malaysia+1

Medical Image Classification Pseudo-Labeling Optical Coherence Tomography Conformal Risk Control

Open paper AskPDF arXiv

The advanced retinal disease diagnosing imaging modality, optical coherence tomography (OCT), encounters a lack of automation because of the high expenses for annotations performed by specialists. The use of SSL solves the problem of insufficient annotations using unlabeled B-scans; however, most of the current techniques for generating pseudo-labels are based on prediction confidence without considering the asymmetry between different types of errors. This paper proposes TRIAGE, a risk-controlled semi-supervised framework for OCT scans classification, which uses the concept of a patient-level conformal risk controller with an asymmetric cost matrix. TRIAGE unites three crucial modules: a hierarchical classifier that is capable of working with partially abnormal supervision of the disease subtypes, a patient-grouped conformal risk controller with primal-dual coverage control, and a context-aware Transformer teacher for cross-slice verification. On the dataset from Noor Eye Hospital (16,822 B-scans, 161 patients, and 554 volumes) with a test set of unseen patients, TRIAGE demonstrates 89.66% scan-level accuracy, 0.8805 macro-F1, 0.9641 macro-AUC, and an 8.34% under-grading rate when using only 20% of the labeled data. With only 5% of the labeled data, TRIAGE keeps 76.88% accuracy and a 0.1656 under-grading rate. Compared with the other six state-of-the-art semi-supervised methods, TRIAGE significantly outperforms them with ablation study demonstrating the contribution of each module in the overall framework performance (by 42.7% in terms of under-grading rate comparing to fixed threshold methods). TRIAGE demonstrates 98.00% accuracy for 3-class classification with 1% labeled data and 95.94% accuracy for 8-class classification with 10% labeled data on the OCT-C8 dataset.

Abstract

Explore similar work

TRIAGE: Risk-Controlled Pseudo-Label Admission for Annotation-Efficient Semi-Supervised Retinal OCT Classification

Chronological Contrastive Learning: Few-Shot Progression Assessment in Irreversible Diseases

Pixel-level Counterfactual Contrastive Learning for Medical Image Segmentation