Table 3.

Tabular summary of public and private available AMD datasets.

Dataset	Description	Images	Purpose	Resolution	Annotation	Availability	Modality
AREDS [57]	Longitudinal study monitoring AMD progression	86,770	Classification	Various	Graded by ocular experts	Public	CFP
ARIA [58]	450 JPEG images with AMD, DR, and healthy controls	450	Classification	Various	Annotated by ophthalmologists	Public	CFP
STARE [59]	400 retinal images with manually segmented blood vessels	400	Classification, Segmentation	($700\times 605$)	Expert-labelled abnormalities	Public	CFP
ODIR [60]	Contains 8 retinal diseases (i.e., Diabetes, AMD, myopia, glaucoma, hypertension, cataract, normal, and others) from the Peking University National Institute of Health Sciences	5000	Classification	Various	Annotations by ophthalmologists	Public	CFP
DHU Dataset [20]	45 OCT volumes with normal, DME, and dry AMD from Duke University, Harvard University, and the University of Michigan	45	Classification	Various	Expert-labelled classes	Public	OCT
Optretina [61]	labeled retinal images from telemedicine screening programs.	306,302	Classification	Various	Annotated by ophthalmologists	Public	CFP, OCT
iChallenge-AMD [62]	Fundus image dataset with AMD and non-AMD labels	-	Classification	Various	Reference standard derived from medical records	Public	CFP, OCT
UMN Database [63]	600 OCT B-scan images from the University of Minnesota Ophthalmology Clinic	600	Segmentation	Various	Ophthalmologists	Public	OCT
OPTIMA Database [64]	Cyst segmentation challenge at the MICCAI 2015 conference	15 training, 15 testing volumes	Segmentation	Various	Medical University of Vienna experts	Public	OCT
RETOUCH Database [65]	Retinal fluid (fluid/cyst) labels: IRF, SRF, and PED from the 2017 MICCAI conference’s retinal OCT fluid challenge	70 volumes	Segmentation	Various	Retina specialists	Public	OCT
VRC Database [65]	Eyes of patients afflicted with AMD, DME, and RVO, three disorders known to cause macular fluid from Medical University of Vienna’s Department of Ophthalmology	1200 volume	Segmentation	Various	Retina specialists	Public	OCT
NEH Database [20]	OCT dataset from Noor Eye Hospital, Tehran	148	Classification	3.5 μm axial resolution	Anonymized annotations	Private	OCT
KMC Database [1]	Fundus image dataset from Kasturba Medical College	1110	Classification	$2588\times 1958$	Expert annotations	Private	CFP
KSH Database [67]	High-resolution images from patients with early AMD from Department of Ophthalmology at Kangbuk Samsung Hospital	775	Classification	Various	Expert annotations	Private	CFP