Table 1 Summary of performance metrics for different architectures in maxilla segmentation
From: A deep learning based automated maxillary sinus segmentation and bone grafts analysis in CBCT images
3D V-Net | MS-D network | Res-UNet | 3D U-Net | |
|---|---|---|---|---|
Dice (%) | 93.2 ± 1.34 | 92.6 ± 1.70 | 69.3 ± 3.34 | 85.4 ± 3.67 |
IOU (%) | 92.1 ± 1.70 | 87.4 ± 1.28 | 53.2 ± 2.39 | 78.8 ± 3.34 |
Precision (%) | 93.2 ± 1.41 | 92.8 ± 1.50 | 70.3 ± 3.73 | 88.2 ± 2.98 |
Sensitivity (%) | 92.1 ± 1.61 | 92.5 ± 1.21 | 70.3 ± 2.41 | 86.1 ± 3.44 |
HD95 (mm) | 1.6 ± 1.04 | 2.7 ± 1.84 | 21.1 ± 4.88 | 3.5 ± 0.73 |
ASD (mm) | 0.5 ± 0.37 | 0.7 ± 0.35 | 4.4 ± 1.07 | 0.9 ± 0.21 |
