Table 3 Comparative results of ablation experiments with different modules table.
From: Segmentation of liver CT images based on weighted medical transformer model
Model | F1 score/% | IoU fraction/% | ACC score/% |
|---|---|---|---|
U-Net | 74.12 | 58.43 | 94.36 |
U-Net + Axial Attention | 75.15 | 60.97 | 96.53 |
U-Net + weighted axial attention | 81.55 | 68.97 | 97.31 |
U-Net + fine-grained coarse-grained fusion mechanism | 81.99 | 69.84 | 97.63 |
Weighted medical transformer | 88.48 | 79.41 | 98.49 |
