A novel dataset and benchmark, which features 1482 RGB-D scans of 478 environments across multiple time steps. Each scene includes several objects whose positions change over time, together with ground truth annotations of object instances and their respective 6DoF mappings among re-scans.
Source: RIO: 3D Object Instance Re-Localization in Changing Indoor Environments
Variants: 3RScan, 3R-Scan
This dataset is used in 2 benchmarks:
| Task | Model | Paper | Date |
|---|---|---|---|
| 3D Object Detection | UniDet3D | UniDet3D: Multi-dataset Indoor 3D Object … | 2024-09-06 |
| Point Cloud Registration | SG-PGM | SG-PGM: Partial Graph Matching Network … | 2024-03-28 |
| Point Cloud Registration | SGAligner | SGAligner : 3D Scene Alignment … | 2023-04-28 |
| 3D Object Detection | TR3D | TR3D: Towards Real-Time Indoor 3D … | 2023-02-06 |
| 3D Object Detection | FCAF3D | FCAF3D: Fully Convolutional Anchor-Free 3D … | 2021-12-01 |
Recent papers with results on this dataset: