# U2-Net: A Bayesian U-Net model with epistemic uncertainty feedback for photoreceptor layer segmentation in pathological OCT scans

Source: [U2-Net: A Bayesian U-Net model with epistemic uncertainty feedback for photoreceptor layer segmentation in pathological OCT scans](<https://www.sota2.com/research/paper/68fa66801e209131cea10b12>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-01-03T14:46:09.437Z

Authors: Jos\\'e Ignacio Orlando, Philipp Seeb\\"ock, Hrvoje Bogunovi\\'c, Sophie Klimscha, Christoph Grechenig, Sebastian Waldstein, Bianca S. Gerendas, Ursula Schmidt-Erfurth

Published: 2019-01-23T14:52:33.000Z

## Abstract

In this paper, we introduce a Bayesian deep learning based model for segmenting the photoreceptor layer in pathological OCT scans. Our architecture provides accurate segmentations of the photoreceptor layer and produces pixel-wise epistemic uncertainty maps that highlight potential areas of pathologies or segmentation errors. We empirically evaluated this approach in two sets of pathological OCT scans of patients with age-related macular degeneration, retinal vein oclussion and diabetic macular edema, improving the performance of the baseline U-Net both in terms of the Dice index and the area under the precision/recall curve. We also observed that the uncertainty estimates were inversely correlated with the model performance, underlying its utility for highlighting areas where manual inspection/correction might be needed.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Disruption Detection on A (AMD (early, CNV), DME, RVO) (test)](<https://www.sota2.com/research/sota/disruption-detection-on-a-amd-early-cnv-dme-rvo-test.md>): Disruption Detection; A (AMD (early, CNV), DME, RVO) (test); AUC: 67.12.

- [Disruption Detection on Late AMD (GA) set B (test)](<https://www.sota2.com/research/sota/disruption-detection-on-late-amd-ga-set-b-test.md>): Disruption Detection; Late AMD (GA) set B (test); AUC: 91.01.

- [Photoreceptor Layer Segmentation on A (AMD (early, CNV), DME, RVO) (test)](<https://www.sota2.com/research/sota/photoreceptor-layer-segmentation-on-a-amd-early-cnv-dme-rvo-test.md>): Photoreceptor Layer Segmentation; A (AMD (early, CNV), DME, RVO) (test); AUC: 96.69.

- [Photoreceptor Layer Segmentation on Late AMD (GA) B (test)](<https://www.sota2.com/research/sota/photoreceptor-layer-segmentation-on-late-amd-ga-b-test.md>): Photoreceptor Layer Segmentation; Late AMD (GA) B (test); AUC: 95.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/1901.07929>)

[PDF](<https://arxiv.org/pdf/1901.07929.pdf>)