Dermoscopic Image Classification Using Deep Belief Learning Network Architecture

Table 5

Comparison of cutting edge architecture models with the proposed DBL network.

	Model	Nonsegmented images		Segmented images
		Accuracy (%)	Error (%)	Accuracy (%)	Error (%)
Dermal dataset (200) images)	DBL model (proposed).	73.7	38.0	90.0	31.0
	DCNN model (AlexNet)	55.6	44.4	62.2	60.1
	CNN model (LeeNet)	50.1	89.0	49.2	87.0
	Probabilistic-based deep learning model	51.4	48.6	61.1	38.9
	Convolutional neural network- (CNN-) based U-Net algorithm	49.3	57.7	66.5	33.5