Deep learning based retinal OCT segmentation

M. Pekala, N. Joshi, Tin Yan Liu, N. M. Bressler, D. Cabrera DeBuc, Philippe Burlina

Research output: Contribution to journalArticle

Abstract

We look at the recent application of deep learning (DL) methods in automated fine-grained segmentation of spectral domain optical coherence tomography (OCT) images of the retina. We describe a new method combining fully convolutional networks (FCN) with Gaussian Processes for post processing. We report performance comparisons between the proposed approach, human clinicians, and other machine learning (ML) such as graph based approaches. The approach is demonstrated on an OCT dataset consisting of mild non-proliferative diabetic retinopathy from the University of Miami. The method is shown to have performance on par with humans, also compares favorably with the other ML methods, and appears to have as small or smaller mean unsigned error (equal to 1.06), versus errors ranging from 1.17 to 1.81 for other methods, and compared with human error of 1.10.

Original languageEnglish (US)
Article number103445
JournalComputers in Biology and Medicine
Volume114
DOIs
StatePublished - Nov 1 2019

Fingerprint

Optical tomography
Optical Coherence Tomography
Learning
Learning systems
Diabetic Retinopathy
Retina
Processing
Deep learning
Machine Learning

Keywords

  • Fully convolutional networks
  • Gaussian process regression
  • Neurodegenerative
  • OCT segmentation
  • Retinal and vascular diseases

ASJC Scopus subject areas

  • Computer Science Applications
  • Health Informatics

Cite this

Deep learning based retinal OCT segmentation. / Pekala, M.; Joshi, N.; Liu, Tin Yan; Bressler, N. M.; DeBuc, D. Cabrera; Burlina, Philippe.

In: Computers in Biology and Medicine, Vol. 114, 103445, 01.11.2019.

Research output: Contribution to journalArticle

Pekala, M. ; Joshi, N. ; Liu, Tin Yan ; Bressler, N. M. ; DeBuc, D. Cabrera ; Burlina, Philippe. / Deep learning based retinal OCT segmentation. In: Computers in Biology and Medicine. 2019 ; Vol. 114.
@article{93d7a919b1da4463bcd8a0a1fde64957,
title = "Deep learning based retinal OCT segmentation",
abstract = "We look at the recent application of deep learning (DL) methods in automated fine-grained segmentation of spectral domain optical coherence tomography (OCT) images of the retina. We describe a new method combining fully convolutional networks (FCN) with Gaussian Processes for post processing. We report performance comparisons between the proposed approach, human clinicians, and other machine learning (ML) such as graph based approaches. The approach is demonstrated on an OCT dataset consisting of mild non-proliferative diabetic retinopathy from the University of Miami. The method is shown to have performance on par with humans, also compares favorably with the other ML methods, and appears to have as small or smaller mean unsigned error (equal to 1.06), versus errors ranging from 1.17 to 1.81 for other methods, and compared with human error of 1.10.",
keywords = "Fully convolutional networks, Gaussian process regression, Neurodegenerative, OCT segmentation, Retinal and vascular diseases",
author = "M. Pekala and N. Joshi and Liu, {Tin Yan} and Bressler, {N. M.} and DeBuc, {D. Cabrera} and Philippe Burlina",
year = "2019",
month = "11",
day = "1",
doi = "10.1016/j.compbiomed.2019.103445",
language = "English (US)",
volume = "114",
journal = "Computers in Biology and Medicine",
issn = "0010-4825",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Deep learning based retinal OCT segmentation

AU - Pekala, M.

AU - Joshi, N.

AU - Liu, Tin Yan

AU - Bressler, N. M.

AU - DeBuc, D. Cabrera

AU - Burlina, Philippe

PY - 2019/11/1

Y1 - 2019/11/1

N2 - We look at the recent application of deep learning (DL) methods in automated fine-grained segmentation of spectral domain optical coherence tomography (OCT) images of the retina. We describe a new method combining fully convolutional networks (FCN) with Gaussian Processes for post processing. We report performance comparisons between the proposed approach, human clinicians, and other machine learning (ML) such as graph based approaches. The approach is demonstrated on an OCT dataset consisting of mild non-proliferative diabetic retinopathy from the University of Miami. The method is shown to have performance on par with humans, also compares favorably with the other ML methods, and appears to have as small or smaller mean unsigned error (equal to 1.06), versus errors ranging from 1.17 to 1.81 for other methods, and compared with human error of 1.10.

AB - We look at the recent application of deep learning (DL) methods in automated fine-grained segmentation of spectral domain optical coherence tomography (OCT) images of the retina. We describe a new method combining fully convolutional networks (FCN) with Gaussian Processes for post processing. We report performance comparisons between the proposed approach, human clinicians, and other machine learning (ML) such as graph based approaches. The approach is demonstrated on an OCT dataset consisting of mild non-proliferative diabetic retinopathy from the University of Miami. The method is shown to have performance on par with humans, also compares favorably with the other ML methods, and appears to have as small or smaller mean unsigned error (equal to 1.06), versus errors ranging from 1.17 to 1.81 for other methods, and compared with human error of 1.10.

KW - Fully convolutional networks

KW - Gaussian process regression

KW - Neurodegenerative

KW - OCT segmentation

KW - Retinal and vascular diseases

UR - http://www.scopus.com/inward/record.url?scp=85072536710&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85072536710&partnerID=8YFLogxK

U2 - 10.1016/j.compbiomed.2019.103445

DO - 10.1016/j.compbiomed.2019.103445

M3 - Article

VL - 114

JO - Computers in Biology and Medicine

JF - Computers in Biology and Medicine

SN - 0010-4825

M1 - 103445

ER -