# BanglaOCT2025: A Population-Specific Fovea-Centric OCT Dataset with Self-Supervised Volumetric Restoration Using Flip-Flop Swin Transformers

## TL;DR

The paper introduces BanglaOCT2025, a new OCT dataset for AMD analysis in the Bengali population, with a novel preprocessing and denoising method using Flip-Flop Swin Transformers.

## Contribution

BanglaOCT2025 is the first clinically validated OCT dataset for the Bengali population, with a fovea-centric preprocessing and self-supervised denoising framework.

## Key findings

- The dataset includes 1585 OCT volumes with expert annotations for AMD subtypes.
- Denoising improved downstream AMD classification accuracy from 69.08% to 99.88%.
- The preprocessing pipeline is robust to retinal tilt and acquisition variability.

## Abstract

Background: Age-related macular degeneration (AMD) is a major cause of vision loss, yet publicly available Optical Coherence Tomography (OCT) datasets lack demographic diversity, particularly from South Asian populations. Existing datasets largely represent Western cohorts, limiting AI generalizability. Moreover, raw OCT volumes contain redundant spatial information and speckle noise, hindering efficient analysis. Methods: We introduce BanglaOCT2025, a retrospective dataset collected from the National Institute of Ophthalmology and Hospital (NIOH), Bangladesh, using Nidek RS-330 Duo 2 and RS-3000 Advance systems. We propose a novel preprocessing pipeline comprising two stages: (1) A constraint-based centroid minimization algorithm automatically localizes the foveal center and extracts a fixed 33-slice macular sub-volume, robust to retinal tilt and acquisition variability; and (2) A self-supervised volumetric denoising module based on a Flip-Flop Swin Transformer (FFSwin) backbone suppresses speckle noise without requiring paired clean reference data. Results: The dataset comprises 1585 OCT volumes (202,880 B-scans), including 857 expert-annotated cases (54 DryAMD, 61 WetAMD, and 742 NonAMD). Denoising quality was evaluated using reference-free volumetric metrics, paired statistical analysis, and blinded clinical review by a retinal specialist, confirming preservation of pathological biomarkers and absence of hallucination. Under a controlled paired evaluation using the same classifier with frozen weights, downstream AMD classification accuracy improved from 69.08% to 99.88%, interpreted as an upper-bound estimate of diagnostic signal recoverability rather than independent generalization. Conclusions: BanglaOCT2025 is the first clinically validated OCT dataset representing the Bengali population and establishes a reproducible fovea-centric volumetric preprocessing and restoration framework for AMD analysis, with future validation across independent and multi-centre test cohorts.