Table 2 CT–US kidney registration performance in the ablation study with individual components removed.
From: A two-step deep learning method for 3DCT-2DUS kidney registration during breathing
Methods | Metric (mm) | CT–CT | CT–US |
|---|---|---|---|
Ours | HD | 3.97 ± 1.37 | 3.80 ± 0.87 |
MCD | 1.15 ± 0.57 | 0.94 ± 0.18 | |
Without MIND loss | HD | 4.43 ± 1.40 | 3.74 ± 0.90 |
MCD | 1.38 ± 0.56 | 0.98 ± 0.23 | |
Without decoder layers | HD | 7.66 ± 2.95 | 4.14 ± 1.15 |
MCD | 2.45 ± 1.10 | 1.02 ± 0.20 | |
Without hierarchical transformation regression | HD | 8.15 ± 3.94 | 4.16 ± 0.97 |
MCD | 2.71 ± 1.48 | 1.05 ± 0.21 | |
Without one cycle transfer learning | HD | 11.59 ± 1.57 | 6.71 ± 1.06 |
MCD | 4.09 ± 0.65 | 2.03 ± 0.29 |
