Table 3.
Absolute error in wall thickness between algorithm and consensus ground truth. The error in millimetres was computed separately for the posterior (Post) and anterior (Ant) wall. The best result in each case is underlined for the anterior and posterior walls. The best overall are marked with an asterisk (*).
| ROBI |
LUMC |
INRIA |
||||
|---|---|---|---|---|---|---|
| Post | Ant | Post | Ant | Post | Ant | |
| Case 1 | 0.38 | 0.12 | 0.20 | 0.35 | 0.41 | 1.14 |
| Case 2 | 0.44 | 0.56 | 0.07 | 0.00 | 1.05 | 0.46 |
| Case 3 | 0.12 | 0.11 | 0.23 | 0.00 | 0.10 | 0.00 |
| Case 4 | 0.19 | 0.14 | 0.00 | 0.19 | 0.29 | 0.19 |
| Case 5 | 0.08 | 0.39 | 0.27 | 0.23 | 0.13 | 0.23 |
| Case 6 | 0.00 | 0.13 | 0.29 | 0.29 | 0.03 | 0.13 |
| Case 7 | 0.52 | 0.10 | 0.00 | 0.27 | 0.12 | 0.00 |
| Case 8 | 0.53 | 0.13 | 0.16 | 0.49 | 0.93 | 0.40 |
| Case 9 | 0.60 | 0.35 | 0.29 | 0.35 | 0.46 | 0.00 |
| Case 10 | 0.11 | 0.29 | 0.24 | 0.42 | 0.83 | 0.42 |
| Median | 0.29 | 0.14* | 0.22* | 0.28 | 0.35 | 0.21 |
| Inter-observer difference | Post = 0.25 mm, | Ant = 0.20 mm | ||||