Table 3.

Performance comparison of the proposed network with state-of-the-art classification models. Bold indicates the best performance while the second-best performance (across each metric) is underlined.

Model	Input size	Sensitivity	Specificity	Precision	F1
ResNet-10172	224 × 224 × 3	0.9583	0.9333	0.9718	0.9650
DenseNet-20145	224 × 224 × 3	0.9444	0.9000	0.9577	0.9510
EfficientNetB446	224 × 224 × 3	0.9583	0.9666	0.9857	0.9718
MobileNetv247	224 × 224 × 3	0.9306	0.9000	0.9571	0.9437
Proposed	224 × 224 × 3	0.9722	0.9666	0.9859	0.9790