| 模型 | Accuracy | Precision | Recall | F1 |
| SE-ResNet50 | 90.26 | 88.15 | 93.02 | 90.52 |
| 改进SE-ResNet | 94.77 | 95.45 | 94.02 | 94.73 |
| VIT | 91.63 | 92.38 | 90.40 | 91.38 |
| SE-ResNet-VIT | 96.44 | 97.06 | 95.77 | 96.41 |