Classification on ModelNet40 original (test)
93.8Overall AccuracyGBNet
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GBNetinput type=coords, #points=1k2019.11 | 93.8 | 91 | |
| RS-CNNinput type=coords + voting, #points=1k2019.11 | 93.6 | — | |
| DGCNNinput type=coords, #points=2k2019.11 | 93.5 | 90.7 | |
| SO-Netinput type=coords + norm, #points=5k2019.11 | 93.4 | 90.8 | |
| DensePointinput type=coords + voting, #points=1k2019.11 | 93.2 | — | |
| RS-CNNinput type=coords, #points=1k2019.11 | 92.9 | — | |
| DGCNNinput type=coords, #points=1k2019.11 | 92.9 | 90.2 | |
| KP-Convinput type=coords, #points=1k2019.11 | 92.9 | — | |
| PointASNLinput type=coords, #points=1k2019.11 | 92.9 | — | |
| DensePointinput type=coords, #points=1k2019.11 | 92.8 | — | |
| SpiderCNNinput type=coords + norm, #points=5k2019.11 | 92.4 | — | |
| PCNNinput type=coords, #points=1k2019.11 | 92.3 | — | |
| PointCNNinput type=coords, #points=1k2019.11 | 92.2 | 88.1 | |
| PointNet++input type=coords + norm, #points=5k2019.11 | 91.9 | — | |
| SO-Netinput type=coords, #points=2k2019.11 | 90.9 | 87.3 | |
| Kd-Netinput type=coords, #points=1k2019.11 | 90.6 | — | |
| RGCNNinput type=coords + norm, #points=1k2019.11 | 90.5 | 87.3 | |
| SCNinput type=coords, #points=1k2019.11 | 90 | 87.6 | |
| PointNetinput type=coords, #points=1k2019.11 | 89.2 | 86 | |
| ECCinput type=coords, #points=1k2019.11 | 87.4 | 83.2 |