Table 10. Registration failure rates for registering a point-cloud target shape with outliers. (Experiment 5).
| Exp. | Outliers | Alg. | Failure Rate (%) by Test Case | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | |||
| 5A-i | 5% | ICP | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| GICP | 0.0 | 0.0 | 0.0 | 0.0 | 0.3 | 0.0 | 0.0 | 0.0 | 0.3 | ||
| Robust ICP | 0.0 | 0.3 | 0.0 | 0.0 | 0.0 | 0.0 | 0.3 | 0.3 | 0.3 | ||
| CPD | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | ||
| IMLP | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | ||
| 5A-ii | 10% | ICP | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| GICP | 0.0 | 1.3 | 1.3 | 0.7 | 1.0 | 0.3 | 0.7 | 1.0 | 1.7 | ||
| Robust ICP | 0.3 | 0.0 | 0.0 | 0.0 | 0.3 | 0.3 | 1.0 | 0.3 | 0.3 | ||
| CPD | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | ||
| IMLP | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | ||
| 5A-iii | 20% | ICP | 0.3 | 0.0 | 0.0 | 0.3 | 0.3 | 0.0 | 0.0 | 0.0 | 0.3 |
| GICP | 5.7 | 7.0 | 6.3 | 5.7 | 8.0 | 9.3 | 5.3 | 4.3 | 8.3 | ||
| Robust ICP | 0.3 | 0.3 | 0.3 | 0.3 | 1.0 | 0.3 | 0.0 | 0.0 | 0.0 | ||
| CPD | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | ||
| IMLP | 0.0 | 0.3 | 0.0 | 1.0 | 0.3 | 1.0 | 1.7 | 0.7 | 0.0 | ||
| 5A-iv | 30% | ICP | 1.3 | 0.3 | 1.0 | 0.7 | 0.3 | 1.0 | 0.3 | 1.0 | 0.3 |
| GICP | 17.0 | 13.3 | 18.7 | 15.7 | 21.3 | 15.3 | 14.3 | 18.0 | 15.7 | ||
| Robust ICP | 4.7 | 2.0 | 3.3 | 2.3 | 3.0 | 4.7 | 1.0 | 1.3 | 1.7 | ||
| CPD | 0.0 | 0.0 | 0.0 | 0.3 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | ||
| IMLP | 4.7 | 3.0 | 5.3 | 3.3 | 6.0 | 3.3 | 3.3 | 3.0 | 4.0 | ||
| 5B-i | 5% | ICP | 0.3 | 1.3 | 0.7 | 0.3 | 0.3 | 0.3 | 1.7 | 0.0 | 0.7 |
| GICP | 1.3 | 3.0 | 1.0 | 1.7 | 1.0 | 2.0 | 1.7 | 1.0 | 1.0 | ||
| Robust ICP | 14.3 | 9.7 | 8.3 | 12.7 | 15.0 | 11.3 | 11.0 | 12.7 | 8.0 | ||
| CPD | 1.0 | 0.3 | 0.7 | 1.7 | 1.7 | 1.0 | 0.7 | 1.3 | 1.7 | ||
| IMLP | 1.0 | 0.3 | 0.3 | 0.3 | 0.0 | 0.7 | 2.0 | 0.7 | 1.3 | ||
| 5B-ii | 10% | ICP | 1.0 | 0.3 | 0.3 | 1.7 | 1.0 | 0.7 | 0.3 | 0.7 | 0.7 |
| GICP | 1.7 | 2.7 | 2.3 | 3.0 | 3.3 | 3.3 | 2.0 | 2.7 | 2.3 | ||
| Robust ICP | 12.7 | 12.3 | 11.3 | 11.0 | 12.0 | 10.7 | 10.7 | 11.3 | 9.3 | ||
| CPD | 1.0 | 1.7 | 1.7 | 0.7 | 1.0 | 1.7 | 1.3 | 0.7 | 0.7 | ||
| IMLP | 1.0 | 0.3 | 1.3 | 2.0 | 0.7 | 1.0 | 1.3 | 0.7 | 1.0 | ||
| 5B-iii | 20% | ICP | 0.7 | 0.7 | 1.0 | 1.3 | 0.7 | 0.0 | 0.3 | 0.7 | 0.7 |
| GICP | 5.0 | 5.0 | 8.0 | 9.0 | 9.7 | 8.3 | 8.7 | 5.3 | 7.3 | ||
| Robust ICP | 10.0 | 7.0 | 9.7 | 13.7 | 11.0 | 14.7 | 12.7 | 10.0 | 11.3 | ||
| CPD | 1.0 | 2.3 | 1.3 | 1.3 | 1.3 | 2.3 | 0.7 | 1.7 | 1.0 | ||
| IMLP | 2.7 | 4.3 | 2.0 | 4.7 | 6.7 | 4.7 | 5.7 | 3.7 | 3.3 | ||
| 5B-iv | 30% | ICP | 2.7 | 1.0 | 1.7 | 1.3 | 1.3 | 1.3 | 2.0 | 2.0 | 0.7 |
| GICP | 15.0 | 17.7 | 12.3 | 17.7 | 19.3 | 20.7 | 18.3 | 21.7 | 20.7 | ||
| Robust ICP | 16.3 | 10.0 | 15.3 | 15.3 | 16.0 | 15.3 | 12.7 | 17.0 | 16.0 | ||
| CPD | 2.0 | 3.3 | 3.3 | 3.7 | 1.7 | 2.0 | 2.0 | 1.7 | 2.3 | ||
| IMLP | 18.7 | 12.0 | 16.0 | 16.7 | 19.0 | 20.7 | 15.3 | 16.7 | 15.7 | ||
Source shapes were randomly generated from a mesh model of a human hip (Fig. 1A), misaligned by [15, 30] mm / degrees in (Experiment 5A) and [30, 60] mm / degrees in (Experiment 5B), and registered back to a point-cloud representation of the mesh. The test cases represent the different noise models used to generate noise on the source shape (Table 4). Outliers were added to the source shape constituting 5% (-i), 10% (-ii), 20% (-iii), and 30% (-iv) of the source points. For each test case, 300 randomized trials were conducted with the percent of unsuccessful registrations (TRE > 10 mm) being shown in the table. The proposed IMLP algorithm was evaluated relative to standard ICP [1], GICP [11], a robust variant of ICP [4], and CPD [20].