| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Scene Retrieval | 3RScan (test) | MRR95 | 16 | |
| 3D Point Cloud Registration | 3RScan (test) | CD0.0083 | 13 | |
| Scene Graph Node Alignment | 3RScan (val) | Mean RR98.7 | 9 | |
| 3D Instance Segmentation | 3RScan ScanNet200→3RScan transfer (test) | AP16.8 | 8 | |
| 3D Object Detection | 3RScan (val) | mAP@0.2568.1 | 8 | |
| 4D Instance Semantic Segmentation | 3RScan v1 (test) | tmAP34.8 | 7 | |
| Scene graph prediction | 3RScan 20 object and 8 predicate classes (test) | Recall (Relationship)68.3 | 6 | |
| 3D Scene Graph Prediction | 3RScan 160 object and 26 predicate classes (test) | Recall (Rel.)68.7 | 6 | |
| 3D Object Detection | 3RScan | mAP@0.2564.7 | 6 | |
| Overlap Check | 3RScan | Precision99.63 | 6 | |
| Temporal Instance Matching (Point-to-Point) | 3RScan v1 (test) | Recall@25%92.31 | 5 | |
| Image-to-Image Scene Retrieval | 3RScan (val) | Temporal Recall@10.1702 | 5 | |
| Point Cloud Registration | 3RScan 65 (test) | RR61.11 | 5 | |
| Instance Reconstruction | 3RScan | L1-Chamfer Distance6.16 | 5 | |
| Cross-Modal Coarse Visual Localization | 3RScan Static Scenario | Recall@1 (K=10)53.6 | 4 | |
| Point-Cloud-to-Point-Cloud Scene Retrieval | 3RScan (val) | Temporal Recall@119.15 | 4 | |
| Relationship-to-Relationship Scene Retrieval | 3RScan (val) | Temporal Recall@119.15 | 4 | |
| Cross-Modal Scene Retrieval (Point Cloud to Description) | 3RScan | Scene Matching Recall@16.71 | 4 | |
| Cross-Modal Scene Retrieval (Image to Description) | 3RScan | Scene Matching Recall @18.72 | 4 | |
| Cross-Modal Scene Retrieval (Image to Point Cloud) | 3RScan | Scene Matching Recall@114.01 | 4 | |
| Instance Matching | 3RScan 65 | Instance Recall (Static)60.32 | 4 | |
| Overlap check | 3RScan (val) | Precision95.41 | 4 | |
| Point cloud mosaicking | 3RScan (143 scenes) | Accuracy12.13 | 4 | |
| Point Cloud Registration | 3RScan (val) | Rotational Error (RRE)1.57 | 4 | |
| 3D Visual Grounding | 3RScan (test) | Unique Success Rate80.4 | 3 |