Table 5. Performance comparison of different attention mechanisms embedded in the V-Net Model for the segmentation task in the LUNA16 dataset.
| Model | DSC | IoU | Precision | Recall |
|---|---|---|---|---|
| V-Net + CA | 0.7694 | 0.6386 | 0.8184 | 0.7539 |
| V-Net + GGCA | 0.7725 | 0.6437 | 0.7577 | 0.8184 |
| V-Net + SE | 0.7656 | 0.6331 | 0.8030 | 0.7529 |
| V-Net + SK | 0.7725 | 0.6446 | 0.8019 | 0.7690 |
| V-Net + GSA | 0.7747 | 0.6426 | 0.7835 | 0.7946 |
CA, coordinate attention; DSC, Dice similarity coefficient; GGCA, global grouped coordinate attention; GSA, grouped split attention; IoU, intersection over union; LUNA16, Lung Nodule Analysis 2016; SE, squeeze and excitation; SK, selective kernel.