Automatic segmentation of nine retinal layer boundaries in OCT images of non-exudative AMD patients using deep learning and graph search

Leyuan Fang, David Cunefare, Chong Wang, Robyn H. Guymer, Shutao Li, Sina Farsiu

Research output: Contribution to journalArticleResearchpeer-review

136 Citations (Scopus)

Abstract

We present a novel framework combining convolutional neural networks (CNN) and graph search methods (termed as CNN-GS) for the automatic segmentation of nine layer boundaries on retinal optical coherence tomography (OCT) images. CNN-GS first utilizes a CNN to extract features of specific retinal layer boundaries and train a corresponding classifier to delineate a pilot estimate of the eight layers. Next, a graph search method uses the probability maps created from the CNN to find the final boundaries. We validated our proposed method on 60 volumes (2915 B-scans) from 20 human eyes with non-exudative age-related macular degeneration (AMD), which attested to effectiveness of our proposed technique.

Original languageEnglish
Pages (from-to)2732-2744
Number of pages13
JournalBiomedical Optics Express
Volume8
Issue number5
DOIs
Publication statusPublished - 1 May 2017
Externally publishedYes

Keywords

  • Image analysis
  • Image processing
  • Optical coherence tomography

Cite this

Fang, Leyuan ; Cunefare, David ; Wang, Chong ; Guymer, Robyn H. ; Li, Shutao ; Farsiu, Sina. / Automatic segmentation of nine retinal layer boundaries in OCT images of non-exudative AMD patients using deep learning and graph search. In: Biomedical Optics Express. 2017 ; Vol. 8, No. 5. pp. 2732-2744.
@article{3b70d6fe0cec47e68db1d6a3c41a8132,
title = "Automatic segmentation of nine retinal layer boundaries in OCT images of non-exudative AMD patients using deep learning and graph search",
abstract = "We present a novel framework combining convolutional neural networks (CNN) and graph search methods (termed as CNN-GS) for the automatic segmentation of nine layer boundaries on retinal optical coherence tomography (OCT) images. CNN-GS first utilizes a CNN to extract features of specific retinal layer boundaries and train a corresponding classifier to delineate a pilot estimate of the eight layers. Next, a graph search method uses the probability maps created from the CNN to find the final boundaries. We validated our proposed method on 60 volumes (2915 B-scans) from 20 human eyes with non-exudative age-related macular degeneration (AMD), which attested to effectiveness of our proposed technique.",
keywords = "Image analysis, Image processing, Optical coherence tomography",
author = "Leyuan Fang and David Cunefare and Chong Wang and Guymer, {Robyn H.} and Shutao Li and Sina Farsiu",
year = "2017",
month = "5",
day = "1",
doi = "10.1364/BOE.8.002732",
language = "English",
volume = "8",
pages = "2732--2744",
journal = "Biomedical Optics Express",
issn = "2156-7085",
publisher = "Optical Society of America (OSA)",
number = "5",

}

Automatic segmentation of nine retinal layer boundaries in OCT images of non-exudative AMD patients using deep learning and graph search. / Fang, Leyuan; Cunefare, David; Wang, Chong; Guymer, Robyn H.; Li, Shutao; Farsiu, Sina.

In: Biomedical Optics Express, Vol. 8, No. 5, 01.05.2017, p. 2732-2744.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Automatic segmentation of nine retinal layer boundaries in OCT images of non-exudative AMD patients using deep learning and graph search

AU - Fang, Leyuan

AU - Cunefare, David

AU - Wang, Chong

AU - Guymer, Robyn H.

AU - Li, Shutao

AU - Farsiu, Sina

PY - 2017/5/1

Y1 - 2017/5/1

N2 - We present a novel framework combining convolutional neural networks (CNN) and graph search methods (termed as CNN-GS) for the automatic segmentation of nine layer boundaries on retinal optical coherence tomography (OCT) images. CNN-GS first utilizes a CNN to extract features of specific retinal layer boundaries and train a corresponding classifier to delineate a pilot estimate of the eight layers. Next, a graph search method uses the probability maps created from the CNN to find the final boundaries. We validated our proposed method on 60 volumes (2915 B-scans) from 20 human eyes with non-exudative age-related macular degeneration (AMD), which attested to effectiveness of our proposed technique.

AB - We present a novel framework combining convolutional neural networks (CNN) and graph search methods (termed as CNN-GS) for the automatic segmentation of nine layer boundaries on retinal optical coherence tomography (OCT) images. CNN-GS first utilizes a CNN to extract features of specific retinal layer boundaries and train a corresponding classifier to delineate a pilot estimate of the eight layers. Next, a graph search method uses the probability maps created from the CNN to find the final boundaries. We validated our proposed method on 60 volumes (2915 B-scans) from 20 human eyes with non-exudative age-related macular degeneration (AMD), which attested to effectiveness of our proposed technique.

KW - Image analysis

KW - Image processing

KW - Optical coherence tomography

UR - http://www.scopus.com/inward/record.url?scp=85019034945&partnerID=8YFLogxK

U2 - 10.1364/BOE.8.002732

DO - 10.1364/BOE.8.002732

M3 - Article

VL - 8

SP - 2732

EP - 2744

JO - Biomedical Optics Express

JF - Biomedical Optics Express

SN - 2156-7085

IS - 5

ER -