Table 3 Accuracy, precision, recall, F1-score, and specificity results obtained for each model and each class separately.

Model	Class	Accuracy (\(\%\))	Precision (\(\%\))	Recall (\(\%\))	F1-score (\(\%\))	Specificity (\(\%\))
EfficientNet	AVL	93.33	73.68	93.33	82.35	95.65
	DRUSEN	77.19	91.67	77.19	83.81	94.52
	NORMAL	91.38	84.13	91.38	87.60	86.11
InceptionV3	AVL	93.33	100.00	93.33	96.55	99.13
	DRUSEN	88.52	93.10	88.52	90.76	94.20
	NORMAL	94.44	87.93	94.44	91.07	90.80
ResNet50	AVL	93.33	93.33	93.33	93.33	99.13
	DRUSEN	87.72	94.34	87.72	90.91	95.89
	NORMAL	96.55	90.32	96.55	93.33	91.67
VGG16	AVL	80.00	92.31	80.00	85.71	86.14
	DRUSEN	85.96	90.74	85.96	88.29	93.15
	NORMAL	93.10	85.71	93.10	89.26	87.50
ViT	AVL	93.33	87.50	93.33	90.32	98.26
	DRUSEN	89.47	91.07	89.47	90.27	93.15
	NORMAL	91.38	91.38	91.38	91.38	93.06
RS-A ViT	AVL	93.33	93.33	93.33	93.33	99.13
	DRUSEN	94.74	98.18	94.74	96.43	98.63
	NORMAL	100.00	96.67	100.00	98.31	97.22

Quick links

Search