Object Classification on ModelNet10
95.5Class AccSO-Net (3-layer)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SO-Net (3-layer)Representation=points + normal, Input=5000 x 6, Pre-training=false2018.03 | 95.5 | 95.7 | |
| SO-Net (2-layer)Representation=points + normal, Input=5000 x 6, Pre-training=true2018.03 | 94.9 | 95 | |
| SO-Net (2-layer)Representation=points + normal, Input=5000 x 6, Pre-training=false2018.03 | 94.4 | 94.5 | |
| SO-Net (2-layer)Representation=points, Input=2048 x 3, Pre-training=false2018.03 | 93.9 | 94.1 | |
| Kd-NetRepresentation=points, Input=2^15 x 32018.03 | 93.5 | 94 | |
| OctNetRepresentation=octree, Input=128^32018.03 | 90.1 | 90.9 | |
| ECCRepresentation=points, Input=1000 x 32018.03 | 90 | 90.8 |