# Anatomy-Aware Unsupervised Detection and Localization of Retinal Abnormalities in Optical Coherence Tomography

arXiv: https://arxiv.org/abs/2604.22139 · HTML: https://mlsanity.com/paper/2604.22139/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 17 of its 38 references that are in the map.
- First posted on arXiv 24 Apr 2026.

## Builds on (most cited first, 17 of 17)

- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Neural Discrete Representation Learning](https://mlsanity.com/paper/1711.00937/) (2017)
- [A ConvNet for the 2020s](https://mlsanity.com/paper/2201.03545/) (2022)
- [An Introduction to Variational Autoencoders](https://mlsanity.com/paper/1906.02691/) (2019)
- [Unsupervised Anomaly Detection with Generative Adversarial Networks to Guide Marker Discovery](https://mlsanity.com/paper/1703.05921/) (2017)
- [MAD-GAN: Multivariate Anomaly Detection for Time Series Data with Generative Adversarial Networks](https://mlsanity.com/paper/1901.04997/) (2019)
- [Autoencoders for Unsupervised Anomaly Segmentation in Brain MR Images: A Comparative Study](https://mlsanity.com/paper/2004.03271/) (2020)
- [Encoding Structure-Texture Relation with P-Net for Anomaly Detection in Retinal Images](https://mlsanity.com/paper/2008.03632/) (2020)
- [Anomaly detection through latent space restoration using vector-quantized variational autoencoders](https://mlsanity.com/paper/2012.06765/) (2020)
- [Anomaly Detection in Medical Imaging -- A Mini Review](https://mlsanity.com/paper/2108.11986/) (2021)
- [Exploiting Epistemic Uncertainty of Anatomy Segmentation for Anomaly Detection in Retinal OCT](https://mlsanity.com/paper/1905.12806/) (2019)
- [Anomaly Detection in Medical Imaging with Deep Perceptual Autoencoders](https://mlsanity.com/paper/2006.13265/) (2020)
- [MedIAnomaly: A comparative study of anomaly detection in medical images](https://mlsanity.com/paper/2404.04518/) (2024)
- [Unsupervised Identification of Disease Marker Candidates in Retinal OCT Imaging Data](https://mlsanity.com/paper/1810.13404/) (2018)
- [Anomaly Detection in Retinal Images using Multi-Scale Deep Feature Sparse Coding](https://mlsanity.com/paper/2201.11506/) (2022)
- [HealthyGAN: Learning from Unannotated Medical Images to Detect Anomalies Associated with Human Disease](https://mlsanity.com/paper/2209.01822/) (2022)
- [Unsupervised Anomaly Detection using Aggregated Normative Diffusion](https://mlsanity.com/paper/2312.01904/) (2023)

## Abstract (arXiv)

Reliable automated analysis of Optical Coherence Tomography (OCT) imaging is crucial for diagnosing retinal disorders but faces a critical barrier: the need for expensive, labor-intensive expert annotations. Supervised deep learning models struggle to generalize across diverse pathologies, imaging devices, and patient populations due to their restricted vocabulary of annotated abnormalities. We propose an unsupervised anomaly detection framework that learns the normative distribution of healthy retinal anatomy without lesion annotations, directly addressing annotation efficiency challenges in clinical deployment. Our approach leverages a discrete latent model trained on normal B-scans to capture OCT-specific structural patterns. To enhance clinical robustness, we incorporate retinal layer-aware supervision and structured triplet learning to separate healthy from pathological representations, improving model reliability across varied imaging conditions. During inference, anomalies are detected and localized via reconstruction discrepancies, enabling both image and pixel-level identification without requiring disease-specific labels. On the Kermany dataset (AUROC: 0.799), our method substantially outperforms VAE, VQVAE, VQGAN, and f-AnoGAN baselines. Critically, cross-dataset evaluation on Srinivasan achieves AUROC 0.884 with superior generalization, demonstrating robust domain adaptation. On the external RETOUCH benchmark, unsupervised anomaly segmentation achieves competitive Dice (0.200) and mIoU (0.117) scores, validating reproducibility across institutions.