DeshadowGAN: A Deep Learning Approach to Remove Shadows from Optical Coherence Tomography Images.
deep learning
generative adversarial network
glaucoma
shadow removal
Journal
Translational vision science & technology
ISSN: 2164-2591
Titre abrégé: Transl Vis Sci Technol
Pays: United States
ID NLM: 101595919
Informations de publication
Date de publication:
04 2020
04 2020
Historique:
received:
30
09
2019
accepted:
13
12
2019
entrez:
21
8
2020
pubmed:
21
8
2020
medline:
21
8
2020
Statut:
epublish
Résumé
To remove blood vessel shadows from optical coherence tomography (OCT) images of the optic nerve head (ONH). Volume scans consisting of 97 horizontal B-scans were acquired through the center of the ONH using a commercial OCT device for both eyes of 13 subjects. A custom generative adversarial network (named DeshadowGAN) was designed and trained with 2328 B-scans in order to remove blood vessel shadows in unseen B-scans. Image quality was assessed qualitatively (for artifacts) and quantitatively using the intralayer contrast-a measure of shadow visibility ranging from 0 (shadow-free) to 1 (strong shadow). This was computed in the retinal nerve fiber layer (RNFL), the inner plexiform layer (IPL), the photoreceptor (PR) layer, and the retinal pigment epithelium (RPE) layer. The performance of DeshadowGAN was also compared with that of compensation, the standard for shadow removal. DeshadowGAN decreased the intralayer contrast in all tissue layers. On average, the intralayer contrast decreased by 33.7 ± 6.81%, 28.8 ± 10.4%, 35.9 ± 13.0%, and 43.0 ± 19.5% for the RNFL, IPL, PR layer, and RPE layer, respectively, indicating successful shadow removal across all depths. Output images were also free from artifacts commonly observed with compensation. DeshadowGAN significantly corrected blood vessel shadows in OCT images of the ONH. Our algorithm may be considered as a preprocessing step to improve the performance of a wide range of algorithms including those currently being used for OCT segmentation, denoising, and classification. DeshadowGAN could be integrated to existing OCT devices to improve the diagnosis and prognosis of ocular pathologies.
Identifiants
pubmed: 32818084
doi: 10.1167/tvst.9.2.23
pii: TVST-19-1971
pmc: PMC7396186
doi:
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Pagination
23Informations de copyright
Copyright 2020 The Authors.
Déclaration de conflit d'intérêts
Disclosure: H. Cheong, None; S.K. Devalla, None; T.H. Pham, None; L. Zhang, None; T.A. Tun, None; X. Wang, None; S. Perera, None; L. Schmetterer, None; T. Aung, None; C. Boote, None; A. Thiery, Abyss Processing (I); M.J.A. Girard, Abyss Processing (I)
Références
Am J Ophthalmol. 1983 May;95(5):673-91
pubmed: 6846459
Arch Ophthal. 1948 Apr;39(4):517-35
pubmed: 18123283
Invest Ophthalmol Vis Sci. 2011 Sep 29;52(10):7738-48
pubmed: 21551412
Sci Rep. 2019 Oct 8;9(1):14454
pubmed: 31595006
Invest Ophthalmol Vis Sci. 2003 Dec;44(12):5189-95
pubmed: 14638716
Transl Vis Sci Technol. 2015 May 15;4(3):3
pubmed: 26046005
Prog Retin Eye Res. 2005 Jan;24(1):39-73
pubmed: 15555526
Invest Ophthalmol Vis Sci. 2013 Mar 01;54(3):2238-47
pubmed: 23449723
Invest Ophthalmol Vis Sci. 2009 Jun;50(6):2785-95
pubmed: 19168906
Opt Express. 2005 Nov 14;13(23):9480-91
pubmed: 19503151
Retina. 2010 Apr;30(4):607-16
pubmed: 20094011
Invest Ophthalmol Vis Sci. 2003 Feb;44(2):623-37
pubmed: 12556392
J Biomed Opt. 2009 Jan-Feb;14(1):010503
pubmed: 19256685
Nat Methods. 2012 Jun 28;9(7):676-82
pubmed: 22743772
IEEE Trans Med Imaging. 2009 Sep;28(9):1436-47
pubmed: 19278927
Opt Lett. 2001 May 1;26(9):608-10
pubmed: 18040398
Science. 2002 Feb 8;295(5557):1077-9
pubmed: 11834836
Nat Methods. 2019 Jan;16(1):67-70
pubmed: 30559429
Can J Ophthalmol. 2006 Feb;41(1):9-12, 14
pubmed: 16462866
Crit Rev Biomed Eng. 1994;22(5-6):401-65
pubmed: 8631195
Ophthalmology. 2010 Dec;117(12):2337-44
pubmed: 20678802
Invest Ophthalmol Vis Sci. 2012 Apr 02;53(4):1714-28
pubmed: 22395883