Towards topological correct segmentation of macular OCT from cascaded FCNs

Yufan He, Aaron Carass, Yeyi Yun, Can Zhao, Bruno M. Jedynak, Sharon Solomon, Shiv Saidha, Peter Calabresi, Jerry Ladd Prince

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Optical coherence tomography (OCT) is used to produce high resolution depth images of the retina and is now the standard of care for in-vivo ophthalmological assessment. In particular, OCT is used to study the changes in layer thickness across various pathologies. The automated image analysis of these OCT images has primarily been performed with graph based methods. Despite the preeminence of graph based methods, deep learning based approaches have begun to appear within the literature. Unfortunately, they cannot currently guarantee the strict biological tissue order found in human retinas. We propose a cascaded fully convolutional network (FCN) framework to segment eight retina layers and preserve the topological relationships between the layers. The first FCN serves as a segmentation network which takes retina images as input and outputs the segmentation probability maps of the layers. We next perform a topology check on the segmentation and those patches that do not satisfy the topology criterion are passed to a second FCN for topology correction. The FCNs have been trained on Heidelberg Spectralis images and validated on both Heidelberg Spectralis and Zeiss Cirrus images.

Original languageEnglish (US)
Title of host publicationFetal, Infant and Ophthalmic Medical Image Analysis - International Workshop, FIFI 2017 and 4th International Workshop, OMIA 2017 Held in Conjunction with MICCAI 2017, Proceedings
PublisherSpringer Verlag
Pages202-209
Number of pages8
Volume10554 LNCS
ISBN (Print)9783319675602
DOIs
StatePublished - 2017
EventInternational Workshop on Fetal and Infant Image Analysis, FIFI 2017 and 4th International Workshop on Ophthalmic Medical Image Analysis, OMIA 2017 held in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017 - Quebec City, Canada
Duration: Sep 14 2017Sep 14 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10554 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

OtherInternational Workshop on Fetal and Infant Image Analysis, FIFI 2017 and 4th International Workshop on Ophthalmic Medical Image Analysis, OMIA 2017 held in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017
CountryCanada
CityQuebec City
Period9/14/179/14/17

Fingerprint

Optical Coherence Tomography
Optical tomography
Retina
Segmentation
Topology
Pathology
Image analysis
Biological Tissue
Tissue
Graph in graph theory
Image Analysis
Patch
High Resolution
Output

Keywords

  • Fully convolutional network
  • Retina OCT
  • Topology preserving

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

He, Y., Carass, A., Yun, Y., Zhao, C., Jedynak, B. M., Solomon, S., ... Prince, J. L. (2017). Towards topological correct segmentation of macular OCT from cascaded FCNs. In Fetal, Infant and Ophthalmic Medical Image Analysis - International Workshop, FIFI 2017 and 4th International Workshop, OMIA 2017 Held in Conjunction with MICCAI 2017, Proceedings (Vol. 10554 LNCS, pp. 202-209). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10554 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-67561-9_23

Towards topological correct segmentation of macular OCT from cascaded FCNs. / He, Yufan; Carass, Aaron; Yun, Yeyi; Zhao, Can; Jedynak, Bruno M.; Solomon, Sharon; Saidha, Shiv; Calabresi, Peter; Prince, Jerry Ladd.

Fetal, Infant and Ophthalmic Medical Image Analysis - International Workshop, FIFI 2017 and 4th International Workshop, OMIA 2017 Held in Conjunction with MICCAI 2017, Proceedings. Vol. 10554 LNCS Springer Verlag, 2017. p. 202-209 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10554 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

He, Y, Carass, A, Yun, Y, Zhao, C, Jedynak, BM, Solomon, S, Saidha, S, Calabresi, P & Prince, JL 2017, Towards topological correct segmentation of macular OCT from cascaded FCNs. in Fetal, Infant and Ophthalmic Medical Image Analysis - International Workshop, FIFI 2017 and 4th International Workshop, OMIA 2017 Held in Conjunction with MICCAI 2017, Proceedings. vol. 10554 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10554 LNCS, Springer Verlag, pp. 202-209, International Workshop on Fetal and Infant Image Analysis, FIFI 2017 and 4th International Workshop on Ophthalmic Medical Image Analysis, OMIA 2017 held in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017, Quebec City, Canada, 9/14/17. https://doi.org/10.1007/978-3-319-67561-9_23
He Y, Carass A, Yun Y, Zhao C, Jedynak BM, Solomon S et al. Towards topological correct segmentation of macular OCT from cascaded FCNs. In Fetal, Infant and Ophthalmic Medical Image Analysis - International Workshop, FIFI 2017 and 4th International Workshop, OMIA 2017 Held in Conjunction with MICCAI 2017, Proceedings. Vol. 10554 LNCS. Springer Verlag. 2017. p. 202-209. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-67561-9_23
He, Yufan ; Carass, Aaron ; Yun, Yeyi ; Zhao, Can ; Jedynak, Bruno M. ; Solomon, Sharon ; Saidha, Shiv ; Calabresi, Peter ; Prince, Jerry Ladd. / Towards topological correct segmentation of macular OCT from cascaded FCNs. Fetal, Infant and Ophthalmic Medical Image Analysis - International Workshop, FIFI 2017 and 4th International Workshop, OMIA 2017 Held in Conjunction with MICCAI 2017, Proceedings. Vol. 10554 LNCS Springer Verlag, 2017. pp. 202-209 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{75fdc3126ba74f66b5daafbf64b11d8a,
title = "Towards topological correct segmentation of macular OCT from cascaded FCNs",
abstract = "Optical coherence tomography (OCT) is used to produce high resolution depth images of the retina and is now the standard of care for in-vivo ophthalmological assessment. In particular, OCT is used to study the changes in layer thickness across various pathologies. The automated image analysis of these OCT images has primarily been performed with graph based methods. Despite the preeminence of graph based methods, deep learning based approaches have begun to appear within the literature. Unfortunately, they cannot currently guarantee the strict biological tissue order found in human retinas. We propose a cascaded fully convolutional network (FCN) framework to segment eight retina layers and preserve the topological relationships between the layers. The first FCN serves as a segmentation network which takes retina images as input and outputs the segmentation probability maps of the layers. We next perform a topology check on the segmentation and those patches that do not satisfy the topology criterion are passed to a second FCN for topology correction. The FCNs have been trained on Heidelberg Spectralis images and validated on both Heidelberg Spectralis and Zeiss Cirrus images.",
keywords = "Fully convolutional network, Retina OCT, Topology preserving",
author = "Yufan He and Aaron Carass and Yeyi Yun and Can Zhao and Jedynak, {Bruno M.} and Sharon Solomon and Shiv Saidha and Peter Calabresi and Prince, {Jerry Ladd}",
year = "2017",
doi = "10.1007/978-3-319-67561-9_23",
language = "English (US)",
isbn = "9783319675602",
volume = "10554 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "202--209",
booktitle = "Fetal, Infant and Ophthalmic Medical Image Analysis - International Workshop, FIFI 2017 and 4th International Workshop, OMIA 2017 Held in Conjunction with MICCAI 2017, Proceedings",

}

TY - GEN

T1 - Towards topological correct segmentation of macular OCT from cascaded FCNs

AU - He, Yufan

AU - Carass, Aaron

AU - Yun, Yeyi

AU - Zhao, Can

AU - Jedynak, Bruno M.

AU - Solomon, Sharon

AU - Saidha, Shiv

AU - Calabresi, Peter

AU - Prince, Jerry Ladd

PY - 2017

Y1 - 2017

N2 - Optical coherence tomography (OCT) is used to produce high resolution depth images of the retina and is now the standard of care for in-vivo ophthalmological assessment. In particular, OCT is used to study the changes in layer thickness across various pathologies. The automated image analysis of these OCT images has primarily been performed with graph based methods. Despite the preeminence of graph based methods, deep learning based approaches have begun to appear within the literature. Unfortunately, they cannot currently guarantee the strict biological tissue order found in human retinas. We propose a cascaded fully convolutional network (FCN) framework to segment eight retina layers and preserve the topological relationships between the layers. The first FCN serves as a segmentation network which takes retina images as input and outputs the segmentation probability maps of the layers. We next perform a topology check on the segmentation and those patches that do not satisfy the topology criterion are passed to a second FCN for topology correction. The FCNs have been trained on Heidelberg Spectralis images and validated on both Heidelberg Spectralis and Zeiss Cirrus images.

AB - Optical coherence tomography (OCT) is used to produce high resolution depth images of the retina and is now the standard of care for in-vivo ophthalmological assessment. In particular, OCT is used to study the changes in layer thickness across various pathologies. The automated image analysis of these OCT images has primarily been performed with graph based methods. Despite the preeminence of graph based methods, deep learning based approaches have begun to appear within the literature. Unfortunately, they cannot currently guarantee the strict biological tissue order found in human retinas. We propose a cascaded fully convolutional network (FCN) framework to segment eight retina layers and preserve the topological relationships between the layers. The first FCN serves as a segmentation network which takes retina images as input and outputs the segmentation probability maps of the layers. We next perform a topology check on the segmentation and those patches that do not satisfy the topology criterion are passed to a second FCN for topology correction. The FCNs have been trained on Heidelberg Spectralis images and validated on both Heidelberg Spectralis and Zeiss Cirrus images.

KW - Fully convolutional network

KW - Retina OCT

KW - Topology preserving

UR - http://www.scopus.com/inward/record.url?scp=85029787070&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85029787070&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-67561-9_23

DO - 10.1007/978-3-319-67561-9_23

M3 - Conference contribution

AN - SCOPUS:85029787070

SN - 9783319675602

VL - 10554 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 202

EP - 209

BT - Fetal, Infant and Ophthalmic Medical Image Analysis - International Workshop, FIFI 2017 and 4th International Workshop, OMIA 2017 Held in Conjunction with MICCAI 2017, Proceedings

PB - Springer Verlag

ER -