[Retracted] BID-Net: An Automated System for Bone Invasion Detection Occurring at Stage T4 in Oral Squamous Carcinoma Using Deep Learning

Table 5

Effect of learning rate on model’s performance.

Learning models	Learning rate	Accuracy	F1-score	Precision	Recall	AUC-score
BID-net	0.01	90.6	89.12	87.68	91.32	94.58
	0.001	91.27	90.16	88.39	93.86	93.24
	0.000 1	93.62	92.63	91	95.17	95.9
	0.000 01	80.2	72.85	77.77	70.88	93.37
VGG-16	0.01	87.87	84.62	85.33	83.39	86.7
	0.001	85.95	82.12	82.9	81.44	87.29
	0.000 1	83.19	78.33	79.51	77.4	84.42
VGG-19	0.000 01	70.79	57.88	60.86	57.56	72.04
	0.01	84.47	79.78	79.3	80.33	86.01
	0.001	84.84	80.46	81.73	79.46	85.64
	0.000 1	78.78	70.54	74.35	68.87	83.23
	0.000 01	68.87	51.8	55.22	52.88	70.47
ResNet-50	0.01	83.47	79.66	79.33	80.03	84.82
	0.001	79.33	70.83	75.58	68.95	86.87
	0.000 1	84.29	79.88	80.86	79.08	89.95
	0.000 01	85.67	82.27	82.08	82.47	90.1
MobileNet	0.01	84.16	79.23	80.45	79	88.95
	0.001	81.54	76.35	77.22	75.65	85.59
	0.000 1	82.92	77.43	79.6	75.99	89
	0.000 01	76.3	64.89	71.1	63.5	80.64
DenseNet121	0.01	82.65	82.01	83.58	81.13	89.18
	0.001	83.98	83.61	83.31	84.18	91.37
	0.000 1	84.43	75.3	83.42	72.1	94.5
	0.000 01	78.51	78.61	78.31	79.12	90.12
ResNet101	0.01	79.25	76.37	75.32	82.69	88.9
	0.001	78.09	69.13	70.31	68.28	89.45
	0.000 1	76.08	74.18	75.17	83.17	87.21
	0.000 01	80.4	77.02	75.54	81.9	88.5