Point Cloud Registration On 3Dlomatch 10 30
Métriques
Recall ( correspondence RMSE below 0.2)
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | Recall ( correspondence RMSE below 0.2) |
|---|---|
| regtr-end-to-end-point-cloud-correspondences | 64.8 |
| the-perfect-match-3d-point-cloud-matching | 33 |
| pcam-product-of-cross-attention-matrices-for | 54.9 |
| deep-global-registration | 48.7 |
| fully-convolutional-geometric-features | 40.1 |
| d3feat-joint-learning-of-dense-detection-and | 37.2 |
| predator-registration-of-3d-point-clouds-with | 62.5 |
| omnet-learning-overlapping-mask-for-partial | 8.4 |
| neighborhood-aware-geometric-encoding-network | 71.9 |
| geometric-transformer-for-fast-and-robust | 74 |
| predator-registration-of-3d-point-clouds-with | 59.8 |
| predator-registration-of-3d-point-clouds-with | 24 |