Semantic Segmentation on S3DIS (Area 2)
80.9mIoUJOPP-3D
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| JOPP-3DSetting=Open Vocab, Supervision=weakly-supervised2026.03 | 80.9 | — | 87 | |
| ConcertoSetting=Closed Vocab, Supervision=supervised2026.03 | 77.4 | — | 85 | |
| SonataSetting=Closed Vocab, Supervision=supervised2026.03 | 76 | — | 81.6 | |
| PointTransformerV3Setting=Closed Vocab, Supervision=supervised2026.03 | 73.4 | — | 78.9 | |
| PonderV2Pre-training strategy=PonderV22023.10 | 71.7 | 79.9 | 91.8 | |
| PointTransformerV2Setting=Closed Vocab, Supervision=supervised2026.03 | 71.6 | — | 77.9 | |
| PointTransformerSetting=Closed Vocab, Supervision=supervised2026.03 | 70.4 | — | 76.5 | |
| PointContrastSetting=Closed Vocab, Supervision=supervised2026.03 | 68.1 | — | 73.5 | |
| KPConvSetting=Closed Vocab, Supervision=supervised2026.03 | 67.1 | — | 72.8 | |
| PPTPre-training strategy=PPT2023.10 | 65.4 | 75.6 | 88.3 | |
| JOPP-3DSetting=Open Vocab, Supervision=unsupervised2026.03 | 59.4 | — | 70.1 | |
| DGCNN (VoxelSSL)Pre-training=VoxelSSL, Few-shot Setting=5-way 10-shot, Backbone=DGCNN2020.09 | 57.73 | 60.45 | — | |
| DGCNN (Ours)Pre-training=Self-supervised (Ours), Few-shot Setting=5-way 10-shot, Backbone=DGCNN2020.09 | 56.43 | 64.67 | — | |
| DGCNN (Random Init)Pre-training=None, Few-shot Setting=5-way 10-shot, Backbone=DGCNN2020.09 | 55.94 | 61.52 | — | |
| OpenMask3DSetting=Open Vocab, Supervision=weakly-supervised2026.03 | 36.7 | — | 43.6 |