Table 3 Detection performance of recent models on LUNA16 and Tianchi
From: GLANCE: continuous global-local exchange with consensus fusion for robust nodule segmentation
Model | Task | LUNA16 detection | Tianchi detection | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Â | Â | Pre. | Recall | F1 | FP | Pre. | Recall | F1 | FP |
GLANCE (ours) | Seg+Det | 92.3 | 94.1 | 93.2 | 7.7 | 95.0 | 93.6 | 94.3 | 5.0 |
LN-DETR34 | Det | 90.5 | 91.9 | 91.2 | 9.5 | 89.5 | 90.7 | 90.1 | 10.5 |
Santone et al.21 | Seg+Det | 91.2 | 91.5 | 91.3 | 8.8 | 90.1 | 90.9 | 90.5 | 9.9 |
NDLA20 | Det | 92.0 | 90.7 | 91.3 | 8.0 | 91.5 | 90.1 | 90.8 | 8.5 |
Shah et al.35 | Det | 91.8 | 90.1 | 90.9 | 8.2 | 89.9 | 88.5 | 89.2 | 10.1 |
Sui et al.36 | Seg+Det | 87.2 | 89.8 | 88.5 | 12.8 | 85.5 | 88.1 | 86.8 | 14.5 |
UrRehman et al.37 | Det | 89.5 | 88.7 | 89.1 | 10.5 | 87.8 | 87.0 | 87.4 | 12.2 |
Moturi et al.38 | Det | 84.5 | 88.2 | 86.3 | 15.5 | 82.9 | 86.5 | 84.7 | 17.1 |
Alhajim et al.26 | Seg+Det | 88.1 | 87.9 | 88.0 | 11.9 | 87.0 | 86.8 | 86.9 | 13.0 |
AWEU-Net27 | Seg+Det | 90.1 | 87.5 | 88.8 | 9.9 | 88.9 | 86.2 | 87.5 | 11.1 |
Lung-CADex39 | Det | 88.5 | 86.0 | 87.2 | 11.5 | 88.2 | 85.8 | 87.0 | 11.8 |
CSE-GAN40 | Det | 86.9 | 85.5 | 86.2 | 13.1 | 84.5 | 83.9 | 84.2 | 15.5 |
EHO-Deep CNN41 | Det | 86.2 | 85.1 | 85.6 | 13.8 | 84.0 | 83.3 | 83.6 | 16.0 |
NHNN42 | Det | 75.3 | 72.1 | 73.7 | 24.7 | 73.8 | 70.5 | 72.1 | 26.2 |
DeepSEED43 | Det | 87.1 | 88.2 | 87.6 | 12.9 | 86.7 | 87.8 | 87.2 | 13.3 |
3D RPN44 | Det | 91.0 | 91.6 | 91.3 | 9.0 | 93.2 | 92.4 | 92.8 | 6.8 |
Lu et al.45 | Det | 86.0 | 87.5 | 86.7 | 14.0 | 85.6 | 86.4 | 86.0 | 14.4 |
NoduleNet46 | Seg+Det | 88.3 | 87.4 | 87.8 | 11.7 | 86.1 | 86.8 | 86.4 | 13.9 |
