Table 6.

Compared with the state-of-the-art methods for binary classification, the best results are in bold.

Models	Accuracy (%)	Sensitivity (%)	Specificity (%)	AUC(%)
MC-CNN [17]	87.14	77.00	93.00	93.00
ResNet50 + SVM-RBF [18]	88.41	85.38	–	93.19
MoDenseNet [19]	90.40	90.47	–	95.48
Deep local–global Network [20]	88.46	88.66	–	95.62
Multi-view CNN [21]	89.90	85.26	90.66	94.85
Paper [22]	81.73	78.24	85.97	–
This work	92.19	92.10	91.50	97.05