Table 10 Comparison of accuracy, precision, recall, and F1 score among different models
From: Reunion helper: an edge matcher for sibling fragment identification of the Dunhuang manuscript
Models | Accuracy (%) | Precision (%) | Recall (%) | F1 score (%) |
|---|---|---|---|---|
AlexNet | 93.697 | 93.800 | 93.697 | 93.611 |
VGG11 | 89.163 | 89.886 | 89.163 | 88.749 |
ResNet18 | 95.217 | 95.293 | 95.217 | 95.166 |
DenseNet121 | 94.450 | 94.485 | 94.450 | 94.398 |
SqueezeNet1.0 | 65.066 | 42.335 | 65.066 | 51.295 |
MobileNetV1 | 93.140 | 93.648 | 93.140 | 93.221 |
MobileNetV2 | 94.401 | 94.518 | 94.401 | 94.326 |
MobileNetV3 | 75.243 | 77.788 | 75.243 | 75.764 |
Ours | 97.630 | 97.668 | 97.630 | 97.638 |