Table 2.

Performance of ViTCNX and the other six models under three classification.

Metrics	Precision	Recall	Accuracy	F1-Score
EfficientNetV2	0.7783	0.4188	0.4526	0.3221
ConvNeXt	0.9562	0.9397	0.9574	0.9473
DenseNet	0.9487	0.9402	0.9560	0.9442
Swin Transformer	0.9259	0.8754	0.9127	0.8957
ResNet-50	0.9369	0.8936	0.9317	0.9100
Vision Transformer	0.9657	0.9599	0.9689	0.9627
ViTCNX	0.9668	0.9597	0.9696	0.9631

Bold values means the highest score under this metric.