Table 1.

The used abbreviations

ACC	Accuracy	MA	Microaneurysms
AUC	Area Under Curve	ME	Macular Edema
AP	Average Pooling	MLC	Multi-Label Classification
APTOS 2019	Asia Pacific Tele-Ophthalmology Society	MLSVM	ML support vector machine
B-HM	Blot Hemorrhages	MP	max-pooling
BV	Blood Vessels	MSE	Mean Squared Error
BPs	bifurcation points	NPDR	Non-proliferative DR
CAD	Computer-Aided Diagnostic	NV	Neovascularization
CM	confusion matrix	OC	optic cup
CNN	Convolutional Neural Network	OCT	Optical Coherence Tomography
CONV	convolution	OCTA	OCT Angiography
CLAHE	contrast limited adaptive histogram equalization
D-HM	Dot Hemorrhages	OD	Optic Disc
DL	Deep Learning	ON	Optic Nerve
DO	dropout	PA	Padding
DR	Diabetic Retinopathy	PDR	Proliferative DR
DSC	Dice Similarity Coefficient	PO	Pooling
EX	Exudates	QKS	Quadratic Kappa Score
FC	fully connected	ReLU	Rectified-Linear-Unit
F-HM	Flame Hemorrhages	RESNET	Residential Energy Services Network
FKM	fuzzy k-means	ROC	Receiver Operating Characteristic
FN	False Negative	ROIs	region of interest
FOV	Field of View	S	Stride
FP	False positive	S-EX	Soft Exudates
FRCNN	Fast Region-based CNN	SEN	Sensitivity
GANs	generative adversarial networks	SGD	Stochastic Gradient Descent
GAP	Global Average Pooling	SHAP	Shapley Additive exPlanations
GT	Ground Truth	SPE	Specificity
H-EX	Hard Exudates	TN	True Negative
HEBPDS	Histogram Equalization for Brightness Preservation Based on a Dynamic Stretching Technique	TP	True positive
HM	Hemorrhages	VGG	Very Deep Convolutional Networks
IDRiD	Indian diabetic retinopathy image dataset	VL	Venous Loops
$L_r$	Learning Rate	VR	Venous Reduplication