Research map / 3D vision / 3D generation
OpenScene: 3D Scene Understanding With Open Vocabularies
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use OpenScene instead of MSeg Voting | ScanNet unseen classes (4 classes) · ScanNet unseen classes (4 classes) | mIoU: 53.4 -> 62.8 | Helped |
| Use OpenScene instead of MSeg Voting | nuScenes semantic segmentation · nuScenes validation | mIoU: 31.0 -> 42.1 | Helped |
| Use OpenScene instead of MSeg Voting | ScanNet semantic segmentation · ScanNet validation | mIoU: 45.6 -> 54.2 | Helped |
| Use OpenScene instead of MSeg Voting | Matterport3D semantic segmentation · Matterport3D test | mIoU: 33.4 -> 43.4 | Helped |
| Increase the label set from 21 to 40 classes | OpenScene versus fully supervised segmentation · Matterport3D, 40 classes | mAcc: 50.8 -> 50.9 | Mixed |
| Use the 2D-3D ensemble instead of 2D fusion only (ablation) | OpenScene with LSeg features · ScanNet validation | mIoU: 50.0 -> 54.2 | Helped |
| Use the 2D-3D ensemble instead of 3D distillation only (ablation) | OpenScene with LSeg features · ScanNet validation | mIoU: 52.9 -> 54.2 | Helped |
| Use the 2D-3D ensemble instead of 2D fusion only (ablation) | OpenScene with LSeg features · Matterport3D test | mIoU: 32.3 -> 43.4 | Helped |
| Use the 2D-3D ensemble instead of 3D distillation only (ablation) | OpenScene with LSeg features · Matterport3D test | mIoU: 41.9 -> 43.4 | Helped |
| Use the 2D-3D ensemble instead of 2D fusion only (ablation) | OpenScene with OpenSeg features · ScanNet validation | mIoU: 41.4 -> 47.5 | Helped |
| Use the 2D-3D ensemble instead of 3D distillation only (ablation) | OpenScene with OpenSeg features · ScanNet validation | mIoU: 46.0 -> 47.5 | Helped |
| Use the 2D-3D ensemble instead of 2D fusion only (ablation) | OpenScene with OpenSeg features · Matterport3D test | mIoU: 32.4 -> 42.6 | Helped |
Closest papers by meaning
- OpenLex3D: A Tiered Benchmark for Open-Vocabulary 3D Scene Representations (NeurIPS 2025 · similarity 0.92)
- A Unified Framework for 3D Scene Understanding (NeurIPS 2024 · similarity 0.90)
- Masked Point-Entity Contrast for Open-Vocabulary 3D Scene Understanding (CVPR 2025 · similarity 0.90)
- SceneFun3D: Fine-Grained Functionality and Affordance Understanding in 3D Scenes (CVPR 2024 · similarity 0.89)
- Open3DSG: Open-Vocabulary 3D Scene Graphs from Point Clouds with Queryable Objects and Open-Set Relationships (CVPR 2024 · similarity 0.89)
- Open-Vocabulary 3D Semantic Segmentation with Foundation Models (CVPR 2024 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.