Language-based Localization on SQA3D (test)
42.6Accuracy @ 0.5mLoc3R-VLM
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Loc3R-VLM3D input=false, 2D input=true2026.03 | 42.6 | 75.9 | 38.4 | 63 | |
| SIG3D3D input=true, 2D input=false2026.03 | 27.4 | 59.1 | 28.7 | — | |
| View2Cap3D input=true, 2D input=false2026.03 | 17.4 | 36.9 | 24.1 | 28.5 | |
| SIG3D3D input=true, 2D input=false, reproduction=reproduced by [63]2026.03 | 16.8 | 35.2 | 23.4 | 26.3 | |
| 3D-VisTA3D input=true, 2D input=false2026.03 | 11.7 | 34.5 | 16.9 | 24.2 | |
| SQA3D3D input=true, 2D input=false, separate=true2026.03 | 10.3 | 31.4 | 17.1 | 22.8 | |
| SQA3D3D input=true, 2D input=false2026.03 | 9.5 | 29.6 | 8.7 | 16.5 | |
| Random3D input=false, 2D input=false2026.03 | 7.2 | 25.8 | 8.4 | 16.9 |