Open-set Point Cloud Classification on ScanObjectNN (Split2 plus Split1)
88.8Sample AccuracyUSS+MSP
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| USS+MSPBackbone=PointNet++2022.12 | 88.8 | 87.8 | 81.3 | 30.9 | 72.2 | |
| MSPBackbone=PointNet++2022.12 | 88.2 | 87.1 | 80.6 | 30.6 | 72.2 | |
| MLSBackbone=PointNet++2022.12 | 88.2 | 87.1 | 80.2 | 32.3 | 71.4 | |
| USS+MSPBackbone=DGCNN2022.12 | 88.2 | 86.2 | 83.7 | 29.4 | 75.5 | |
| PointCaM (USS+USE)Backbone=PointNet++2022.12 | 88.1 | 86.8 | 79.6 | 30.2 | 73.1 | |
| MSPBackbone=DGCNN2022.12 | 88 | 86.2 | 84.3 | 29.9 | 75.3 | |
| MLSBackbone=DGCNN2022.12 | 88 | 86.2 | 79.3 | 30.3 | 75.5 | |
| PointCaM (USS+USE)Backbone=DGCNN2022.12 | 87.6 | 85.9 | 82.9 | 29.7 | 76 | |
| MSPBackbone=AdaptConv2022.12 | 87.5 | 85.5 | 86.7 | 31.1 | 73.4 | |
| MLSBackbone=AdaptConv2022.12 | 87.5 | 85.5 | 85 | 30.5 | 74.4 | |
| PointCaM (USS+USE)Backbone=AdaptConv2022.12 | 87.1 | 85.5 | 84.1 | 30.1 | 74.4 | |
| USS+MSPBackbone=AdaptConv2022.12 | 86.8 | 84.6 | 84.6 | 30.4 | 74.1 | |
| MSPBackbone=PointNet2022.12 | 80.2 | 75.7 | 86.8 | 39.7 | 63.6 | |
| MLSBackbone=PointNet2022.12 | 80.2 | 75.7 | 89.9 | 42.2 | 60.4 | |
| PointCaM (USS+USE)Backbone=PointNet2022.12 | 79.6 | 75.6 | 86 | 38.6 | 65.5 | |
| USS+MSPBackbone=PointNet2022.12 | 79.3 | 75.5 | 90.5 | 40.5 | 63.3 |