Segmentation of laser induced retinal lesions using deep learning (December 2021).
deep learning
fundus images
laser safety
laser tissue interaction
transfer learning
Journal
Lasers in surgery and medicine
ISSN: 1096-9101
Titre abrégé: Lasers Surg Med
Pays: United States
ID NLM: 8007168
Informations de publication
Date de publication:
10 2022
10 2022
Historique:
revised:
18
05
2022
received:
04
01
2022
accepted:
13
06
2022
pubmed:
6
7
2022
medline:
14
9
2022
entrez:
5
7
2022
Statut:
ppublish
Résumé
Detection of retinal laser lesions is necessary in both the evaluation of the extent of damage from high power laser sources, and in validating treatments involving the placement of laser lesions. However, such lesions are difficult to detect using Color Fundus cameras alone. Deep learning-based segmentation can remedy this, by highlighting potential lesions in the image. A unique database of images collected at the Air Force Research Laboratory over the past 30 years was used to train deep learning models for classifying images with lesions and for subsequent segmentation. We investigate whether transferring weights from models that learned classification would improve performance of the segmentation models. We use Pearson's correlation coefficient between the initial and final training phases to reveal how the networks are transferring features. The segmentation models are able to effectively segment a broad range of lesions and imaging conditions. Deep learning-based segmentation of lesions can effectively highlight laser lesions, making this a useful tool for aiding clinicians.
Identifiants
pubmed: 35781887
doi: 10.1002/lsm.23578
pmc: PMC9464686
mid: NIHMS1816646
doi:
Types de publication
Journal Article
Research Support, N.I.H., Extramural
Research Support, Non-U.S. Gov't
Research Support, U.S. Gov't, Non-P.H.S.
Langues
eng
Sous-ensembles de citation
IM
Pagination
1130-1142Subventions
Organisme : NIGMS NIH HHS
ID : R01 GM127696
Pays : United States
Organisme : NIGMS NIH HHS
ID : R21 GM142107
Pays : United States
Organisme : NCI NIH HHS
ID : R21 CA269099
Pays : United States
Informations de copyright
© 2022 Wiley Periodicals LLC.
Références
J Med Syst. 2018 Oct 31;42(12):247
pubmed: 30382410
Ophthalmol Ther. 2020 Jun;9(2):249-263
pubmed: 32141037
Nature. 2017 Feb 2;542(7639):115-118
pubmed: 28117445
Am J Ophthalmol. 2012 Jan;153(1):137-42.e2
pubmed: 21937017
J Anat. 2005 Apr;206(4):319-48
pubmed: 15817102
Arch Ophthalmol. 1995 Mar;113(3):325-32
pubmed: 7887846
Ophthalmology. 2014 Apr;121(4):946-51
pubmed: 24411577
IEEE Trans Neural Netw Learn Syst. 2016 Nov;27(11):2187-2200
pubmed: 26353384
JAMA Ophthalmol. 2018 Jul 1;136(7):803-810
pubmed: 29801159
IEEE J Biomed Health Inform. 2017 Jan;21(1):76-84
pubmed: 28114048
Surv Ophthalmol. 2000 May-Jun;44(6):459-78
pubmed: 10906379
IEEE Trans Med Imaging. 2015 May;34(5):1018-30
pubmed: 25376036
Dtsch Arztebl Int. 2017 Dec 8;114(49):831-837
pubmed: 29271340
Bull Soc Belge Ophtalmol. 1981;197:9-17
pubmed: 6182936
Aviat Space Environ Med. 2003 Sep;74(9):947-52
pubmed: 14503672
Lasers Surg Med. 2009 Jul;41(5):382-90
pubmed: 19533764
Annu Int Conf IEEE Eng Med Biol Soc. 2015 Aug;2015:3029-32
pubmed: 26736930
Proc Natl Acad Sci U S A. 2017 Mar 28;114(13):3521-3526
pubmed: 28292907
IEEE Trans Med Imaging. 2016 May;35(5):1285-98
pubmed: 26886976
Proc Natl Acad Sci U S A. 2020 Dec 1;117(48):30088-30095
pubmed: 32393633
Nat Med. 2018 Sep;24(9):1342-1350
pubmed: 30104768