Research Article
Recognition of Point Sets Objects in Realistic Scenes
Table 3
Comparisons of classification accuracy (%) on ModelNet40.
| Method | Input | ModelNet40 (%) |
| FPNN [8] | Volume | 68.20 | 3DShapeNets [22] | Volume | 77.30 | VoxNet [4] | Volume | 83.00 | Subvolume [1] | Volume | 86.00 | PointNet(vanilla) [10] | Point cloud | 87.20 | PointNet [10] | Point cloud | 89.20 | This paper | Point cloud | 89.70 |
|
|