| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| ScanObjectNN OBJ_BG (test) | PointGST | Top-1 Accuracy94.49 | 35 | 1mo ago | |
| ShapeNet | Ours w/ PointBind | Accuracy99.59 | 20 | 1mo ago | |
| ScanObjectNN PB_T50_RS fine-tuned | RECON++-L | Overall Accuracy95.25 | 19 | 1mo ago | |
| ModelNet40 1k points (test) | Point2Vec | Accuracy94.8 | 18 | 1mo ago | |
| OmniObject3D | O-Shape + BayesMM | Accuracy (1024 pts)37.77 | 16 | 25d ago | |
| Objaverse-LVIS | Uni3D | Accuracy55.42 | 16 | 25d ago | |
| ScanObjectNN | RECON++-L | Top-1 Accuracy0.654 | 16 | 1mo ago | |
| ScanObjectNN OBJ_ONLY fine-tuned | RECON++-L | Overall Accuracy97.59 | 15 | 1mo ago | |
| ScanObjectNN OBJ_BG fine-tuned | RECON++-L | Overall Accuracy98.8 | 15 | 1mo ago | |
| ScanObjectNN PB_T50_RS (test) | SUGAR (single) | Top-1 Accuracy49.3 | 14 | 1mo ago | |
| Objaverse | SAGE-13B | GPT-4 Score58.48 | 12 | 1mo ago | |
| 3D Objects | GPT-4V | Recognition Accuracy61.33 | 11 | 1mo ago | |
| ModelNet40 8k points (test) | RECON++-L | Accuracy95 | 11 | 1mo ago | |
| SUN RGB-D (common classes split) | CLIP2 | Top-1 Accuracy (Average)69.6 | 6 | 1mo ago | |
| S-OBJ_ONLY base-to-new | ULIP-2 + Point-PRC | Base Accuracy83.6 | 6 | 1mo ago | |
| S-PB_T50_RS base-to-new | ULIP-2 + Point-PRC | Base Score73.67 | 6 | 1mo ago | |
| ScanNet | CLIP² | Mean Top1 Accuracy38.5 | 5 | 1mo ago | |
| ONCE | CLIP2 | Top-1 Accuracy37.8 | 5 | 1mo ago | |
| nuScenes | CLIP2 | Average Top-1 Accuracy37.8 | 5 | 1mo ago | |
| ScanNet 249 classes v2 (test) | CLIP2 | Top-5 Accuracy31.7 | 3 | 1mo ago | |
| ScanNet 384 classes v2 (test) | CLIP2 | Top-5 Accuracy22 | 3 | 1mo ago | |
| SUN RGB-D (test) | - | Top-1 Accuracy (Overall)- | 0 | 1mo ago | |
| ScanNet 249 classes | - | Top5 Acc- | 0 | 1mo ago | |
| ScanNet 384 classes | - | Top-5 Accuracy- | 0 | 1mo ago | |
| ScanNet v2 (test) | - | Mean Top1 Acc- | 0 | 1mo ago |
