Table 3: Performance comparison of different methods (percent).

MethodAccuracyRecallPrecisionSpecificityF1 scoreAUCCI

CNN-F95.7387.1498.1099.3092.3098.49 (0.0029)[97.91 99.06]
CNN-194.9888.6693.9197.6191.2197.83 (0.0033)[97.18 98.49]
CNN-293.1484.1091.8796.9087.8195.63 (0.0052)[94.60 96.66]
ResNet5093.2882.2094.1797.8987.7897.41 (0.003)[97.26 98.28]
InceptionV393.7887.2991.1596.4889.1897.31 (0.0054)[96.85 98.46]
VGG-1692.7983.9090.8396.4887.2296.82 (0.0046)[96.33 97.31]
Peng et al. [11]88.8082.1091.7093.3095.30
Liu et al. [12]86.7391.3082.3582.9691.05

Number format of AUC: mean (standard deviation).