Table 4 Registration results for scenario 2
Scenario 2 | Recall\(_C\) [%] | \(S_{Vg}\) | \(S_{Vs}\) | RMSD | |
|---|---|---|---|---|---|
Ground truth | 100 | 0.420 | 0.672 | 0.195 | |
Feature-based | SIFT + FPFH | 3.3 | 0.993 | 0.990 | 4.924 |
SIFT + PFHRGB | 4.8 | 0.990 | 0.986 | 4.877 | |
SIFT + RIFT | 4 | 0.992 | 0.988 | 4.903 | |
Deep learning | DCP | 14.2 | 0.973 | 0.958 | 4.559 |
PointNetLK | 13.1 | 0.969 | 0.961 | 4.563 | |
DeepGMR | 68.9 | 0.851 | 0.794 | 2.397 | |
GeoTransformer | 90.5 | 0.795 | 0.725 | 0.823 | |
Predator | 94.2 | 0.781 | 0.714 | 0.867 | |
Proposed solution | 96.3 | 0.440 | 0.727 | 0.388 | |