Table 5.

Experimental results using the combined GoogleNet and ResNet18 features with different classifiers.

Method	Accuracy	Precision	Recall	F1-Measure	Training Time
Random forest	97.93	0.979	0.969	0.999	241
Support vector machine	99.90	0.999	0.998	0.999	12,382
Fast decision tree	94.45	0.944	0.915	0.984	302
Bayesian network	81.88	0.814	0.736	0.931	404