Articles Vol. 63 No. 6 (2022) 09/11/2022

SCREENING GLAUCOMA IN COLOR FUNDUS IMAGES USING ARTIFICIAL INTELLIGENCE SOFWARE EYEDR

Pham Thi Thuy Tien1,2, Ho Phuong Thanh Tai1,2, Do Quoc Hiep1,2, Nguyen Trong Loc1,2, Nguyen Minh Son3,4
1 Ho Chi Minh City Eye Hospital
2 Bệnh viện Mắt TPHCM
3 University of Information Technology, Vietnam National University
4 Đại học Công nghệ thông tin - Đại học Quốc Gia TPHCM
Corresponding author: taiho@bvmat.com
DOI: 10.52163/yhc.v63i6.465
21 Views
0 Downloads
Keywords
Glaucoma, optic disc, atrificial intelligence.
Abstract

Using Darknet convolutional neural network technology, YOLO version 3 develops an artificial
intelligence system called EYEDR, that can screen glaucoma by disc imaging automatically.
Purposes: Determining the validity and reliability of the artificial intelligence algorithm EYEDR in
the screening of adult glaucoma by color images of the optic disc.
Methods: Cross-sectional study. The 2470 color images of the optic disc were selected to train
machine learning with Darknet YOLO convolutional neural network to determine the shape of the
optic disc and optic cup, thereby an artificial intelligence algorithm was developed to detect the
glaucoma and non-glaucoma optic nerve. A test dataset of 1028 disc images was used to evaluate
the value of this AI software based on the sensitivity, specificity, area under the ROC curve (AUC),
likelihood ratio, accuracy, and Cohen’s Kappa coefficient.
Results: The area under the curve (AUC) was 0.93 ± 0.01 (95% CI; 0.92 - 0.95), with the sensitivity
90.03%, specificity 95.06%, the accuracy 86.8% and the Kappa coefficient 0.86 with p<0.001.
Conclusions: EYEDR artificial intelligence software developed by CNN Darknet YOLO vesion 3
technology with high sensitivity could screen glaucoma by color fundus images in the community.

References
[1]
Ahn JM, Kim S, Ahn KS et al., “A deep learning Google Scholar
[2]
model for the detection of both advanced and Google Scholar
[3]
early glaucoma using fundus photography” PLOS Google Scholar
[4]
ONE, 1-8, 2018. Google Scholar
[5]
Chakrabarty L, Gopal DJGD, Arunava CA et Google Scholar
[6]
al., Automated Detection of Glaucoma From Google Scholar
[7]
Topographic Features of the Optic Nerve Head Google Scholar
[8]
in Color Fundus Photographs. J Glaucoma. Google Scholar
[9]
(7):590-597, 2016.151 Google Scholar
[10]
Chen X, Xu Y, Wong KW et al., “Glaucoma Google Scholar
[11]
detection based on deep convolutional neural Google Scholar
[12]
network”. IEEE Eng Med Biol Soc, pp. 715 – Google Scholar
[13]
, Annual Conf, 2015. Google Scholar
[14]
Fingeret M, Medeiros FA, Susanna R et al., Five Google Scholar
[15]
rules to evaluate the optic disc and retinal nerve Google Scholar
[16]
fiber layer for Glaucoma. Optometry.;76:661-8, Google Scholar
[17]
Kauppi T, Kalesnykiene V, Kamarainen JK et Google Scholar
[18]
al., “DIARETDB1 diabetic retinopathy database Google Scholar
[19]
and evaluation protocol”. In: Medical image Google Scholar
[20]
understanding and analysis, processdings of 11th Google Scholar
[21]
conference; 2007. Google Scholar
[22]
Li Z, He Y, Keel S et al., “Efficacy of a Deep Google Scholar
[23]
Learning System for Detecting Glaucomatous Google Scholar
[24]
Optic Neuropathy Based on Color Fundus Google Scholar
[25]
Photographs” American academy of Google Scholar
[26]
Ophthalmology, vol. 125, pp. 1199 – 1206, 2018. Google Scholar
[27]
Liu H, Li L, Wormstone IM et al., “Development Google Scholar
[28]
and Validation of a Deep Learning System to Google Scholar
[29]
Detect Glaucomatous Optic Neuropathy using Google Scholar
[30]
Fundus Photographs” JAMA Ophthalmol. E1- Google Scholar
[31]
E8, 2019. Google Scholar
[32]
Phene S, Dunn RC, Hammel N et al., “Deep Google Scholar
[33]
learning and Glaucoma specialists: The relative Google Scholar
[34]
importance of optic disc features to predict Google Scholar
[35]
Glaucoma referral in fundus photographs”. Google Scholar
[36]
Ophthalmology. 2019;1-13. Google Scholar
[37]
Qaisar A, “Glaucoma-Deep: Detection of Google Scholar
[38]
Glaucoma Eye Disease on Retinal Fundus Images Google Scholar
[39]
using Deep Learning” International Journal of Google Scholar
[40]
Advanced Computer Science and Applications, Google Scholar
[41]
(6): 41-45, 2017. Google Scholar
[42]
Redmon J, Divvala S, Girshick R et al., “You Google Scholar
[43]
Only Look Once: Unified, Real-time Object Google Scholar
[44]
Detection,” in Computer vision and Pattern Google Scholar
[45]
Recognition 2016, 2016. Google Scholar
[46]
Redmon J, Farhadi A, “Yolov3: An incremental Google Scholar
[47]
improvement”. arXiv Operational, 2018. Google Scholar
[48]
Thanh THP, Thuy TPT, Hieu TN et al., “A RealTime Classification Of Glaucoma from Retinal Google Scholar
[49]
Fundus Images Using AI Technology,” 2020 Google Scholar
[50]
International Conference on Advanced Computing Google Scholar
[51]
and Applications (ACOMP), 2020, pp. 114-121, Google Scholar
[52]
doi: 10.1109/ACOMP50827.2020.00024. Google Scholar
[53]
Ting DSW, Pasquale LR, Peng L et al., “Artificial Google Scholar
[54]
intelligence and deep learning in ophthalmology”. Google Scholar
[55]
Br J Ophthalmol;103:167–175, 2019. Google Scholar