Table 7.

Quantitative performance comparison of proposed network with the state-of-the-art deep learning methods. (#Par: Total number of parameters).

Study	Method	#Par (Million)	Dataset 1: CT					Dataset 2: X-ray
			ACC	F1	AP	AR	AUC	ACC	F1	AP	AR	AUC
Minaee et al. [16]	SqueezeNet	1.24	89.84	89.48	89.91	89.06	93.86	93.51	93.56	93.43	93.70	97.03
Brunese et al. [13]	VGG16	134.27	89.66	89.54	91.43	87.81	92.35	95.79	95.91	95.65	96.18	97.79
Khan et al. [17]	VGG19	139.58	91.54	91.33	92.26	90.47	94.54	95.30	95.39	95.14	95.65	97.84
Martínez et al. [25]	NASNet	4.27	93.68	93.49	94.19	92.82	96.67	94.06	94.06	93.89	94.23	97.07
Misra et al. [24]	ResNet18	11.18	92.96	92.76	93.41	92.14	95.06	95.59	95.69	95.44	95.95	97.79
Farooq et al. [23]	ResNet50	23.54	90.30	90.22	92.17	88.53	92.79	94.73	94.77	94.62	94.92	97.66
Ardakani et al. [20]	ResNet101	42.56	90.30	90.26	92.17	88.64	95.71	94.53	94.58	94.44	94.72	97.19
Jaiswal et al. [19]	DenseNet201	18.11	94.17	94.03	94.63	93.46	97.36	93.41	93.39	93.38	93.39	97.31
Hu et al. [22]	ShuffleNet	0.86	91.65	91.52	92.69	90.44	95.61	94.97	95.03	94.81	95.25	97.53
Apostolopoulos et al. [18]	MobileNetV2	2.24	92.95	92.85	93.81	91.94	96.51	94.65	94.68	94.51	94.85	97.51
Tsiknakis et al. [21]	InceptionV3	21.81	94.57	94.41	94.89	93.94	97.93	95.44	95.53	95.29	95.78	97.52
Proposed	Ensemble-Net	3.16	94.72	94.60	95.22	94.00	97.50	95.83	95.94	95.68	96.20	97.99