Primitive Fitting on Primitive Fitting (In-distribution)
94.4mIoUSPFN
Evaluation results
| Method | Links | |
|---|---|---|
| SPFNSegmentation Module Type=Classification-based2022.03 | 94.4 | |
| AutoGPart (DGCNN)Backbone=DGCNN, Segmentation Module Type=Classification-based2022.03 | 94.2 | |
| HPNet*Segmentation Module Type=Clustering-based2022.03 | 93.9 | |
| MixStyle (DGCNN)Backbone=DGCNN, Training Strategy=MixStyle, Segmentation Module Type=Classification-based2022.03 | 93.7 | |
| DGCNNBackbone=DGCNN, Segmentation Module Type=Classification-based2022.03 | 93.6 | |
| Gradient Surgery (DGCNN)Backbone=DGCNN, Training Strategy=Gradient Surgery, Segmentation Module Type=Classification-based2022.03 | 92.6 | |
| MixStyle (PN++)Backbone=PointNet++, Training Strategy=MixStyle, Segmentation Module Type=Classification-based2022.03 | 92.1 | |
| Gradient Surgery (PN++)Backbone=PointNet++, Training Strategy=Gradient Surgery, Segmentation Module Type=Classification-based2022.03 | 90.7 | |
| AutoGPart (PN++)Backbone=PointNet++, Segmentation Module Type=Classification-based2022.03 | 86.3 | |
| PointNet++Backbone=PointNet++, Segmentation Module Type=Classification-based2022.03 | 81.5 | |
| Meta-learning (DGCNN)Backbone=DGCNN, Training Strategy=Meta-learning, Segmentation Module Type=Classification-based2022.03 | 67.1 | |
| Meta Learning (PN++)Backbone=PointNet++, Training Strategy=Meta-learning, Segmentation Module Type=Classification-based2022.03 | 65 |