Table 5.

Performance results of CNN-ANN hybrid models based on ROI using the GAC method for early diagnosis and differentiation of skin cancer and other skin lesions.

Hybrid-Models	Type of Lesion	AUC %	Sensitivity %	Accuracy %	Precision %	Specificity %
AlexNet-ANN	Scc	89.1	87.2	86.5	88.6	99.6
	Akiec	92.3	91.2	90.8	88.8	99.5
	Bcc	95.2	98.4	97.6	96.9	100
	Bkl	96.7	97.8	98.3	96.1	99.6
	Df	75.1	66.9	66.7	72.7	99.5
	Mel	95.2	93.4	92.6	88.5	96.8
	Nv	97.1	96.1	95.8	97.5	98.1
	Vasc	74.2	67.4	66.7	82.9	100
GoogLeNet-ANN	Scc	82.5	71.2	71.4	67.7	99.2
	Akiec	88.8	73.8	73.6	75.7	99.4
	Bcc	96.7	93.9	93.7	94.4	98.8
	Bkl	96.4	93.2	93.1	93.3	98.5
	Df	67.2	58.4	58.3	57.1	100
	Mel	96.1	93.7	94.2	92.3	97.8
	Nv	98.4	97.1	97.4	97.7	98.2
	Vasc	92.8	91.4	60.8	70.5	100
VGG16-ANN	Scc	77.2	69.4	69	70.2	98.8
	Akiec	81.2	70.2	70.1	74.8	98.6
	Bcc	95.6	94.7	94.9	93.2	99.3
	Bkl	93.1	92.9	92.6	92.7	99.1
	Df	64.5	60.2	60.4	63	100
	Mel	94.2	93.4	93.5	93.6	99.1
	Nv	98.1	97.7	97.5	97.4	96.8
	Vasc	66.6	62.9	62.7	61.5	99.8