Research Article
Multitask Deep Neural Network for the Fully Automatic Measurement of the Angle of Progression
Table 2
Comparison of the results of different methods.
| | Network | Acc (-) | Dice_ALL (-) | Dice_PS (-) | Dice_FH (-) | Dist_L (mm) | Dist_R (mm) | APT mean (STD) (°) | AoP mean (STD) (°) |
| | Unet | 97.1% | 90.1% | 80.0% | 91.5% | 6.948 | 11.414 | 12.496 (18.756) | 6.646 (7.541) | | MT-Unet_A | 97.6% | 91.2% | 85.3% | 91.9% | 8.131 | 9.109 | 9.047 (13.263) | 5.607 (5.725) | | MT-Unet_B | 98.2% | 91.2% | 85.2% | 91.8% | 9.105 | 8.198 | 6.960 (11.089) | 6.696 (6.951) | | MT-Unet_C | 98.1% | 91.9% | 83.7% | 92.8% | 7.486 | 9.051 | 8.813 (6.640) | 5.428 (5.022) | | MT-Unet_D | 98.4% | 91.8% | 85.8% | 92.1% | 6.638 | 5.586 | 6.614 (6.385) | 5.053 (5.158) |
|
|
