表 4. Comparative experiments of different models on the test set.

不同模型在测试集上的对比实验

模型名称	Acc	Pre	Sens	Spec	F1	参数量	计算量
MobileNetV2	80.52%	85.85%	83.18%	75.82%	84.49%	2.23M	426.15M
ConvNeXt	84.10%	94.14%	80.06%	91.21%	86.53%	27.82M	5 818.47M
FasterNet	83.10%	93.70%	78.82%	90.66%	85.62%	2.63M	443.13M
GhostNet	83.50%	89.67%	83.8%	82.97%	86.63%	6.08M	318.47M
EfficientNet	81.31%	94.51%	75.08%	92.31%	83.68%	4.01M	537.34M
LeViT	86.48%	94.70%	83.49%	91.97%	88.74%	7.00M	349.17M
MobileViT	87.87%	94.52%	85.98%	91.21%	90.05%	5.00M	1 785.98M
SlicerViT	90.66%	94.77%	90.36%	91.21%	92.50%	2.16M	279.10M