3D Object Detection on SUN RGB-D V2 (test)
85.7AP (bed)RS-CNN
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RS-CNNbackbone=RS-CNN, structure=VoteNet2019.11 | 85.7 | 47.3 | 64.9 | 77 | 85.2 | 24.9 | 27.9 | 60.5 | 30.4 | 77.4 | 58.1 | |
| PointNet++backbone=PointNet++, structure=VoteNet2019.11 | 84.8 | 51.3 | 68.5 | 77 | 88.8 | 25.3 | 27.3 | 60.6 | 32.2 | 72.2 | 58.8 | |
| Geometric Back-projection Networkbackbone=GBNet, structure=VoteNet2019.11 | 84.8 | 51 | 64.1 | 77 | 88.1 | 24.4 | 31.2 | 62.9 | 32.4 | 77.5 | 59.3 | |
| DGCNNbackbone=DGCNN, structure=VoteNet2019.11 | 84.6 | 46 | 60 | 76.8 | 89.1 | 19.5 | 31.2 | 54.8 | 24.6 | 72.8 | 55.9 | |
| PointNetbackbone=PointNet, structure=VoteNet2019.11 | 46.2 | 23.9 | 28.3 | 52.1 | 38.4 | 9.5 | 1.4 | 4.6 | 3.6 | 17.8 | 22.6 |