3D Visual Grounding on OpenTarget randomly selected 300 samples
46.2Accuracy @ IoU=0.25Ours
Evaluation results
| Method | Links | ||
|---|---|---|---|
| OursOLT=Mask3D [44] + ACE2025.12 | 46.2 | 34.2 | |
| VLM-GrounderOLT=GT, Note=results on randomly selected 300 samples2025.12 | 28.6 | 20.4 | |
| SeqVLMOLT=GT2025.12 | 19.4 | 19.2 | |
| SeeGroundOLT=GT2025.12 | 17.9 | 17.4 | |
| VLM-GrounderOLT=-, Note=results on randomly selected 300 samples2025.12 | 12.3 | 9.6 | |
| GPT4SceneOLT=GT2025.12 | 12.1 | 11.8 |