Fig 1.
Model architecture diagram.
Fig 2.
Feature extraction and enhancement module.
Fig 3.
Main feature selection module.
Fig 4.
Process of the effectiveness scoring network.
Fig 5.
Comparison of registration performances on unseen shapes.
Table 1.
ModelNet40: Registration results for an invisible shape point cloud.
Fig 6.
Comparison of registration performances on unseen categories.
Table 2.
ModelNet40: Registration results for the invisible point cloud categories.
Fig 7.
Comparison of registration performances on unseen shapes with gaussian noise.
Table 3.
ModelNet40: Registration results for an invisible point cloud with gaussian noise.
Table 4.
ModelNet40: Comparison of registration errors at different noise levels.
Table 5.
Running time of different registration algorithms.
Table 6.
Registration performance on the Stanford dataset.
Table 7.
Registration performance with noise on the Stanford dataset.
Fig 8.
Visualization of key feature points. (a) Visualization of key feature points for object 1. (b) Visualization of key feature points for object 2.
Fig 9.
Registration results on the ModelNet40 dataset. (a) Initial point cloud position of object A. (b) Registered point cloud result of object A. (c) Initial point cloud position of bject B. (d) Registered point cloud result of object B.
Fig 10.
Registration results on the Stanford dataset. (a) Initial point cloud position. (b) Registered point cloud result.
Table 8.
Results of different combinations of key components on ModelNet40.