Table 2.

Evaluation of the identification results of different CNN models.

Model	Precision	Recall	F1-score	Accuracy (%)
SqueezeNet25	0.813	0.810	0.810	80.27
AlexNet50	0.830	0.837	0.830	82.65
VGG16_BN51	0.883	0.876	0.877	87.07
ShuffleNetV252	0.891	0.895	0.891	89.12
GoogleNet23	0.918	0.895	0.900	90.14
ResNet5022	0.924	0.927	0.925	91.84
EfficientNet-B026	0.929	0.933	0.929	92.86
EfficientNet-CA	0.948	0.952	0.948	94.22