SceneFun3D
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| 3D functionality segmentation | SceneFun3D (split0 val) | AP5031.24 | 15 | |
| Task Affordance Prediction | SceneFun3D | Success Rate50.68 | 7 | |
| Functionality segmentation | SceneFun3D 1 (val) | mAP8.1 | 6 | |
| 3D Functionality Segmentation | SceneFun3D 1 (train) | mAP7.1 | 6 | |
| 3D functionality segmentation | SceneFun3D split1 (train) | AP5021.32 | 6 | |
| Scene Graph Generation | SceneFun3D | Object Recall @ 3 (R3)81.8 | 5 | |
| 3D motion prediction | SceneFun3D (test) | ADE (Absolute)0.351 | 5 | |
| 3D Functionality Segmentation | SceneFun3D (Split 1 (200 scenes)) | AP5016.3 | 5 | |
| Scene Reconstruction | SceneFun3D | Objects R@390.5 | 5 | |
| Articulation Estimation | SceneFun3D 30 scenes | Joint Accuracy96.04 | 4 | |
| Node Evaluation | SceneFun3D | Objects Score75.5 | 4 | |
| Triplet Retrieval | SceneFun3D 11 | Overall Score50.9 | 4 | |
| Open Vocabulary Instance Segmentation | SceneFun3D | Runtime (per-scene average, s)12 | 3 | |
| Functional Human-scene Interaction | SceneFun3D curated subset | SCS26.07 | 3 | |
| General Human-scene Interaction | SceneFun3D curated subset | SCS0.2542 | 3 | |
| Functionality Segmentation | SceneFun3D | Runtime (s)78 | 2 | |
| Referring Grounding | SceneFun3D spatial referring subset 1.0 (val) | mAP8.11 | 2 | |
| Functional Relation Prediction | SceneFun3D | Node Association Recall@568.3 | 2 |
Showing 18 of 18 rows