Table 4 Effects of frame point choices The result is classification accuracy on ModelNet40 with 2 layers FPAC network.
From: Frame points attention convolution for deep learning on point cloud
Group | a | b | c | d | e | f | g | h |
|---|---|---|---|---|---|---|---|---|
v | 1 | 5 | 9 | 13 | 17 | 25 | 13 | 13 |
l | 0 | 1 | 2 | 3 | 2 | 3 | 2 | 1 |
Share in local (%) | 1.6 | 7.8 | 14.1 | 20.3 | 26.6 | 39.1 | 20.3 | 20.3 |
Acc (%) | 88.0 | 90.1 | 90.8 | 91.1 | 91.4 | 91.4 | 91.2 | 91.0 |
