Automated segmentation of peripapillary retinal boundaries in OCT combining a convolutional neural network and a multi-weights graph search

Pengxiao Zang, Jie Wang, Tristan T. Hormel, Liang Liu, David Huang, Jia Yali

Research output: Contribution to journalArticle

Abstract

Quantitative analysis of the peripapillary retinal layers and capillary plexuses from optical coherence tomography (OCT) and OCT angiography images depend on two segmentation tasks – delineating the boundary of the optic disc and delineating the boundaries between retinal layers. Here, we present a method combining a neural network and graph search to perform these two tasks. A comparison of this novel method’s segmentation of the disc boundary showed good agreement with the ground truth, achieving an overall Dice similarity coefficient of 0.91 ± 0.04 in healthy and glaucomatous eyes. The absolute error of retinal layer boundaries segmentation in the same cases was 4.10 ± 1.25 µm.

Original languageEnglish (US)
Article number368487
Pages (from-to)4340-4352
Number of pages13
JournalBiomedical Optics Express
Volume10
Issue number8
DOIs
StatePublished - Aug 1 2019

Fingerprint

Optical Coherence Tomography
tomography
Weights and Measures
Optic Disk
Angiography
ground truth
angiography
quantitative analysis
boundary layers
optics
coefficients

ASJC Scopus subject areas

  • Biotechnology
  • Atomic and Molecular Physics, and Optics

Cite this

Automated segmentation of peripapillary retinal boundaries in OCT combining a convolutional neural network and a multi-weights graph search. / Zang, Pengxiao; Wang, Jie; Hormel, Tristan T.; Liu, Liang; Huang, David; Yali, Jia.

In: Biomedical Optics Express, Vol. 10, No. 8, 368487, 01.08.2019, p. 4340-4352.

Research output: Contribution to journalArticle

@article{924b0188e0594f4d99c077066c648488,
title = "Automated segmentation of peripapillary retinal boundaries in OCT combining a convolutional neural network and a multi-weights graph search",
abstract = "Quantitative analysis of the peripapillary retinal layers and capillary plexuses from optical coherence tomography (OCT) and OCT angiography images depend on two segmentation tasks – delineating the boundary of the optic disc and delineating the boundaries between retinal layers. Here, we present a method combining a neural network and graph search to perform these two tasks. A comparison of this novel method’s segmentation of the disc boundary showed good agreement with the ground truth, achieving an overall Dice similarity coefficient of 0.91 ± 0.04 in healthy and glaucomatous eyes. The absolute error of retinal layer boundaries segmentation in the same cases was 4.10 ± 1.25 µm.",
author = "Pengxiao Zang and Jie Wang and Hormel, {Tristan T.} and Liang Liu and David Huang and Jia Yali",
year = "2019",
month = "8",
day = "1",
doi = "10.1364/BOE.10.004340",
language = "English (US)",
volume = "10",
pages = "4340--4352",
journal = "Biomedical Optics Express",
issn = "2156-7085",
publisher = "The Optical Society",
number = "8",

}

TY - JOUR

T1 - Automated segmentation of peripapillary retinal boundaries in OCT combining a convolutional neural network and a multi-weights graph search

AU - Zang, Pengxiao

AU - Wang, Jie

AU - Hormel, Tristan T.

AU - Liu, Liang

AU - Huang, David

AU - Yali, Jia

PY - 2019/8/1

Y1 - 2019/8/1

N2 - Quantitative analysis of the peripapillary retinal layers and capillary plexuses from optical coherence tomography (OCT) and OCT angiography images depend on two segmentation tasks – delineating the boundary of the optic disc and delineating the boundaries between retinal layers. Here, we present a method combining a neural network and graph search to perform these two tasks. A comparison of this novel method’s segmentation of the disc boundary showed good agreement with the ground truth, achieving an overall Dice similarity coefficient of 0.91 ± 0.04 in healthy and glaucomatous eyes. The absolute error of retinal layer boundaries segmentation in the same cases was 4.10 ± 1.25 µm.

AB - Quantitative analysis of the peripapillary retinal layers and capillary plexuses from optical coherence tomography (OCT) and OCT angiography images depend on two segmentation tasks – delineating the boundary of the optic disc and delineating the boundaries between retinal layers. Here, we present a method combining a neural network and graph search to perform these two tasks. A comparison of this novel method’s segmentation of the disc boundary showed good agreement with the ground truth, achieving an overall Dice similarity coefficient of 0.91 ± 0.04 in healthy and glaucomatous eyes. The absolute error of retinal layer boundaries segmentation in the same cases was 4.10 ± 1.25 µm.

UR - http://www.scopus.com/inward/record.url?scp=85070833819&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85070833819&partnerID=8YFLogxK

U2 - 10.1364/BOE.10.004340

DO - 10.1364/BOE.10.004340

M3 - Article

VL - 10

SP - 4340

EP - 4352

JO - Biomedical Optics Express

JF - Biomedical Optics Express

SN - 2156-7085

IS - 8

M1 - 368487

ER -