Table 4.
Performance comparison of different models in DPC-Pre.
| Name | Accuracy (%) | Recall (%) | Precision (%) | F1-Score | AUC |
|---|---|---|---|---|---|
| VGGNet (42) | 48.0 | 50.1 | 49.5 | 0.498 | 0.539 |
| ResNet18 (10) | 90.4 | 93.2 | 93.2 | 0.933 | 0.987 |
| ResNet34 (10) | 92.4 | 96.3 | 95.4 | 0.959 | 0.995 |
| ResNet50 (10) | 87.4 | 90.1 | 90.1 | 0.901 | 0.963 |
| ResNet101 (10) | 85.9 | 88.5 | 88.8 | 0.887 | 0.945 |
| DenseNet (32) | 92.9 | 95.8 | 95.9 | 0.959 | 0.995 |
| SENet_18 (36) | 85.9 | 86.6 | 88.5 | 0.876 | 0.937 |
| ShuffleNet (34) | 89.9 | 92.6 | 92.8 | 0.927 | 0.976 |
| MobileNet V2 (33) | 93.9 | 96.8 | 96.8 | 0.969 | 0.995 |
| ECANet_18 (37) | 94.2 | 97.1 | 97.4 | 0.974 | 0.994 |
| HRNet (39) | 81.3 | 83.9 | 83.9 | 0.839 | 0.940 |
| GhostNet (35) | 88.4 | 91.2 | 91.1 | 0.911 | 0.973 |
| RepVGG (30) | 88.4 | 91.1 | 92.2 | 0.917 | 0.987 |
| Vision Transformer (11) | 82.5 | 86.5 | 80.9 | 0.836 | 0.864 |
| CONVADD (15) | 92.9 | 95.9 | 95.9 | 0.959 | 0.989 |
| Efficient B2 (41) | 87.3 | 90.1 | 90.1 | 0.901 | 0.968 |
| Our Proposed | 95.5 | 97.9 | 98.0 | 0.979 | 0.997 |