Skin-Net: a novel deep residual network for skin lesions classification using multilevel feature extraction and cross-channel correlation with detection of outlier

Journal of Big Data

Table 7 Comparison of performance measures using ISIC-2019 with state-of-the-art

	Method	Augmentation	Classification	Pre-processing (enhancement and segmentation)	Performance measures
	Method	Augmentation	Classification	Pre-processing (enhancement and segmentation)	Accuracy (%)	Specificity (%)	Sensitivity (%)	Precision (%)	F-Score (%)
[59]	Transfer learning GoogleNet, ResNet-101, and NasNet-Larg	All classes	Binary	No	88.33	88.24	88.46	–	–
[56]	Transfer learning & GoogleNet	All classes	Multiclass (8)	Yes	92.99	96	70.44	62.78	66.39
Proposed methods	RDNN	Classes < 1000 image	Multiclass (9)	No	94.65	96.78	70.78	72.56	71.33

The proposed method for ISIC 2019 obtained the highest values for all measures of accuracy, specificity, sensitivity, precision, and F-Score compared with methods [56, 59]