3D Visual Grounding on Nr3D (test)
76.1Overall Success RatePanoGrounder
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| PanoGrounderTraining Regime=Multi Dataset (ScanRefer + ReferIt3D)2025.12 | 76.1 | 84.1 | 68.4 | 72.9 | 77.5 | — | |
| PanoGrounderTraining Regime=Single Dataset2025.12 | 74.6 | 82.2 | 67.2 | 70.5 | 76.3 | — | |
| Transcrib3DEvaluation Protocol=Zero-shot (use gt object class)2026.02 | 70.2 | 79.7 | 60.3 | 60.1 | 75.4 | — | |
| ViewSRDTraining Regime=Single Dataset2025.12 | 69.9 | 75.3 | 64.8 | 68.6 | 70.6 | — | |
| VGMambaTraining Regime=Single Dataset2025.12 | 68.3 | — | 61.4 | — | — | — | |
| EaSeEvaluation Protocol=Zero-shot (use gt object class)2026.02 | 67.8 | — | — | — | — | — | |
| PQ3DTraining setup=joint training2024.05 | 66.7 | 75 | 58.7 | — | — | — | |
| MPECEvaluation Protocol=Fully supervised2026.02 | 66.7 | — | — | — | — | — | |
| PQ3DTraining Regime=Multi Dataset2025.12 | 66.7 | 75 | 58.7 | 62.8 | 68.6 | — | |
| UniVLGTraining Regime=Multi Dataset2025.12 | 65.2 | 73.3 | 57 | 55.1 | 69.9 | — | |
| PQ3DTraining setup=single dataset2024.05 | 64.9 | 73.3 | 56.7 | — | — | — | |
| SceneVerseEvaluation Protocol=Fully supervised2026.02 | 64.9 | 72.5 | 57.8 | 56.9 | 67.9 | — | |
| PQ3D (sg.)Training Regime=Single Dataset2025.12 | 64.9 | 73.3 | 56.7 | 60.7 | 67 | — | |
| GPSTraining Regime=Multi Dataset2025.12 | 64.9 | 72.5 | 57.8 | 56.9 | 67.9 | — | |
| LIBATraining Regime=Single Dataset2025.12 | 64.5 | — | 57.2 | 60.3 | — | — | |
| ViL3DRelProposals=ground-truth object proposals2022.11 | 64.4 | 70.2 | 57.4 | 62 | 64.5 | — | |
| ViL3DRel2024.05 | 64.4 | 70.2 | 57.4 | — | — | — | |
| MIKASA2024.03 | 64.4 | 69.7 | 59.4 | 65.4 | 64 | — | |
| MiKASAEvaluation Protocol=Fully supervised2026.02 | 64.4 | 69.7 | 59.4 | 65.4 | 64 | — | |
| ViL3DRelEvaluation Protocol=Fully supervised2026.02 | 64.4 | 70.2 | 57.4 | 62 | 64.5 | — | |
| ViL3DRelTraining Regime=Single Dataset2025.12 | 64.4 | 70.2 | 57.4 | 62 | 64.5 | — | |
| MIKASATraining Regime=Single Dataset2025.12 | 64.4 | 69.7 | 59.4 | 65.4 | 64 | — | |
| 3D-VisTA2024.05 | 64.2 | 72.1 | 56.7 | — | — | — | |
| 3D-VisTATraining Regime=Multi Dataset2025.12 | 64.2 | 72.1 | 56.7 | 61.5 | 65.1 | — | |
| MCLNTraining Regime=Single Dataset2025.12 | 59.8 | — | — | — | — | — | |
| MVTArch.=Ours, Extra=None, Training Setting=Nr3D w/ Sr3D+2022.04 | 59.5 | 67.4 | 52.7 | 59.1 | 60.3 | — | |
| CSVGEvaluation Protocol=Zero-shot (use gt object class)2026.02 | 59.2 | 67.1 | 51.3 | 53 | 62.5 | — | |
| Z3DVLM=Qwen3-VL-235B-Thinking2026.02 | 58.8 | 68.4 | 47.4 | 58.3 | 59.1 | — | |
| GPS (scr.)Training Regime=Single Dataset2025.12 | 58.7 | 67 | 50.9 | 55.8 | 59.8 | — | |
| MVTArch.=Ours, Extra=None, Training Setting=Nr3D w/ Sr3D2022.04 | 58.5 | 65.6 | 51.6 | 56.6 | 59.4 | — | |
| Z3DVLM=Qwen3-VL-30B-Thinking2026.02 | 57.6 | 66.2 | 47.4 | 57.3 | 57.8 | — | |
| 3D-VisTA (scr.)Training Regime=Single Dataset2025.12 | 57.5 | 65.9 | 49.4 | 53.7 | 59.4 | — | |
| Z3DVLM=Qwen2.5-VL-72B2026.02 | 56.8 | 66.9 | 44.7 | 58.3 | 55.8 | — | |
| SATArch.=SAT, Extra=2D assist., Training Setting=Nr3D w/ Sr3D+2022.04 | 56.5 | 64.9 | 48.4 | 54.4 | 57.6 | — | |
| ViewRefer2024.03 | 56 | 63 | 49.7 | 55.1 | 56.8 | — | |
| SPAZERVLM=Qwen2.5-VL-72B2026.02 | 56 | 60.3 | 50.9 | 54.2 | 57.1 | — | |
| MVTArch.=Ours, Extra=None, Training Setting=Nr3D2022.04 | 55.1 | 61.3 | 49.1 | 54.3 | 55.4 | — | |
| Multi-viewProposals=ground-truth object proposals2022.11 | 55.1 | 61.3 | 49.1 | 54.3 | 55.4 | — | |
| MVT2024.03 | 55.1 | 61.3 | 49.1 | 54.3 | 55.4 | — | |
| Z3DEvaluation Protocol=Zero-shot2026.02 | 54.8 | 62.6 | 47.5 | 50.7 | 57.1 | — | |
| BUTD-DETR2024.03 | 54.6 | 60.7 | 48.4 | 46 | 58 | — | |
| BUTD-DETRTraining Regime=Single Dataset2025.12 | 54.6 | 60.7 | 48.4 | 46 | 58 | — | |
| SPAZEREvaluation Protocol=Zero-shot2026.02 | 54.3 | 62.4 | 46.9 | 49.9 | 56.8 | — | |
| Z3DVLM=Qwen3-VL-8B-Thinking2026.02 | 54 | 59.6 | 47.4 | 54.2 | 53.9 | — | |
| SATArch.=SAT, Extra=2D assist., Training Setting=Nr3D w/ Sr3D2022.04 | 53.9 | 61.5 | 46.7 | 52.7 | 54.5 | — | |
| SeqVLMVLM=Doubao-1.5-vision-pro2026.02 | 53.2 | 58.1 | 47.4 | 51 | 54.5 | — | |
| D-LISA2024.10 | 53.1 | 60.2 | 46.2 | 44.3 | 57.4 | — | |
| LaSPEvaluation Protocol=Zero-shot2026.02 | 52.9 | 60.7 | 45.3 | 49.2 | 54.7 | — | |
| EaSeEvaluation Protocol=Zero-shot2026.02 | 52.9 | — | — | — | — | — | |
| 3D-SPSProposals=ground-truth object proposals2022.11 | 51.5 | 58.1 | 45.1 | 48 | 53.2 | — | |
| M3DRef-CLIP2024.05 | 49.4 | 55.6 | 43.4 | — | — | — | |
| M3DRef-CLIP2024.10 | 49.4 | 55.6 | 43.4 | 42.3 | 52.9 | — | |
| SATArch.=SAT, Extra=2D assist., Training Setting=Nr3D2022.04 | 49.2 | 56.3 | 42.4 | 46.9 | 50.4 | — | |
| SATProposals=ground-truth object proposals2022.11 | 49.2 | 56.3 | 42.4 | 46.9 | 50.4 | — | |
| SAT2024.03 | 49.2 | 56.3 | 42.4 | 46.9 | 50.4 | — | |
| TSP3DTraining Regime=Single Dataset2025.12 | 48.7 | — | — | — | — | — | |
| HAM2024.10 | 48.2 | 54.3 | 41.9 | 41.5 | 51.4 | — | |
| TransRefer3DArch.=TransRefer3D, Extra=None, Training Setting=Nr3D w/ Sr3D+2022.04 | 48 | 56.7 | 39.6 | 42.5 | 50.7 | — | |
| VLM-GrounderVLM=GPT-4o2026.02 | 48 | 55.2 | 39.5 | 45.8 | 49.4 | — | |
| TransRefer3DArch.=TransRefer3D, Extra=None, Training Setting=Nr3D w/ Sr3D2022.04 | 47.2 | 55.4 | 39.3 | 40.3 | 50.6 | — | |
| SeeGroundEvaluation Protocol=Zero-shot2026.02 | 46.1 | 54.5 | 38.3 | 42.3 | 48.2 | — | |
| non-SATArch.=non-SAT, Extra=None, Training Setting=Nr3D w/ Sr3D+2022.04 | 45.9 | — | — | — | — | — | |
| SeeGroundVLM=Qwen2-VL-72B2026.02 | 45.2 | 51.5 | 37.7 | 44.8 | 45.5 | — | |
| LanguageReferArch.=LanguageRefer, Extra=None, Training Setting=Nr3D2022.04 | 43.9 | 51 | 36.6 | 41.7 | 45 | — | |
| non-SATArch.=non-SAT, Extra=None, Training Setting=Nr3D w/ Sr3D2022.04 | 43.9 | — | — | — | — | — | |
| LanguageReferProposals=ground-truth object proposals2022.11 | 43.9 | 51 | 36.6 | 41.7 | 45 | — | |
| TransRefer3DArch.=TransRefer3D, Extra=None, Training Setting=Nr3D2022.04 | 42.1 | 48.5 | 36 | 36.5 | 44.9 | — | |
| TransRefer3DProposals=ground-truth object proposals2022.11 | 42.1 | 48.5 | 36 | 36.5 | 44.9 | — | |
| FFL-3DOGArch.=FFL-3DOG, Extra=None, Training Setting=Nr3D2022.04 | 41.7 | 48.2 | 35 | 37.1 | 44.7 | — | |
| FFL-3DOGProposals=ground-truth object proposals2022.11 | 41.7 | 48.2 | 35 | 37.1 | 44.7 | — | |
| FFL-3DOG2024.10 | 41.7 | 48.2 | 35 | 37.1 | 44.7 | — | |
| 3DVG-TransformerArch.=3DVG-Transformer, Extra=None, Training Setting=Nr3D2022.04 | 40.8 | 48.5 | 34.8 | 34.8 | 43.7 | — | |
| 3DVG-TransProposals=ground-truth object proposals2022.11 | 40.8 | 48.5 | 34.8 | 34.8 | 43.7 | — | |
| 3DVG-Transf.2024.03 | 40.8 | 48.5 | 34.8 | 34.8 | 43.7 | — | |
| 3DVG-Trans2024.10 | 40.8 | 48.5 | 34.8 | 34.8 | 43.7 | — | |
| ZSVG3DEvaluation Protocol=Zero-shot2026.02 | 39 | 46.5 | 31.7 | 36.8 | 40 | — | |
| InstanceReferArch.=InstanceRefer, Extra=None, Training Setting=Nr3D2022.04 | 38.8 | 46 | 31.8 | 34.5 | 41.9 | — | |
| InstanceReferProposals=ground-truth object proposals2022.11 | 38.8 | 46 | 31.8 | 34.5 | 41.9 | — | |
| InstanceRefer2024.03 | 38.8 | 46 | 31.8 | 34.5 | 41.9 | — | |
| InstanceRefer2024.10 | 38.8 | 46 | 31.8 | 34.5 | 41.9 | — | |
| ReferIt3DArch.=ReferIt3D, Extra=None, Training Setting=Nr3D w/ Sr3D+2022.04 | 37.6 | 45.4 | 30 | 33.1 | 39.8 | — | |
| TGNNProposals=ground-truth object proposals2022.11 | 37.3 | 44.2 | 30.6 | 35.8 | 38 | — | |
| TGNN2024.10 | 37.3 | 44.2 | 30.6 | 35.8 | 38 | — | |
| ReferIt3DArch.=ReferIt3D, Extra=None, Training Setting=Nr3D w/ Sr3D2022.04 | 37.2 | 44 | 30.6 | 33.3 | 39.1 | — | |
| ReferIt3DArch.=ReferIt3D, Extra=None, Training Setting=Nr3D2022.04 | 35.6 | 43.6 | 27.9 | 32.5 | 37.1 | — | |
| ReferIt3DProposals=ground-truth object proposals2022.11 | 35.6 | 43.6 | 27.9 | 32.5 | 37.1 | — | |
| ReferIt3D2024.03 | 35.6 | 43.6 | 27.9 | 32.5 | 37.1 | — | |
| ScanReferProposals=ground-truth object proposals2022.11 | 34.2 | 41 | 23.5 | 29.9 | 35.4 | — | |
| 3D-SPSVenue=CVPR'22, Pipeline=Two-stage (gt)2025.02 | — | — | — | — | — | 51.5 | |
| 3D-SPSVenue=CVPR'22, Pipeline=Single-stage2025.02 | — | — | — | — | — | 39.2 | |
| BUTD-DETRVenue=ECCV'22, Pipeline=Two-stage (gt)2025.02 | — | — | — | — | — | 54.6 | |
| BUTD-DETRVenue=ECCV'22, Pipeline=Two-stage (det)2025.02 | — | — | — | — | — | 43.3 | |
| BUTD-DETRVenue=ECCV'22, Pipeline=Single-stage2025.02 | — | — | — | — | — | 38.7 | |
| EDAVenue=CVPR'23, Pipeline=Two-stage (gt)2025.02 | — | — | — | — | — | 52.1 | |
| EDAVenue=CVPR'23, Pipeline=Two-stage (det)2025.02 | — | — | — | — | — | 40.7 | |
| EDAVenue=CVPR'23, Pipeline=Single-stage2025.02 | — | — | — | — | — | 40 | |
| G3-LQVenue=CVPR'24, Pipeline=Two-stage (gt)2025.02 | — | — | — | — | — | 58.4 | |
| InstanceReferVenue=ICCV'21, Pipeline=Two-stage (gt)2025.02 | — | — | — | — | — | 38.8 | |
| InstanceReferVenue=ICCV'21, Pipeline=Two-stage (det)2025.02 | — | — | — | — | — | 29.9 | |
| LanguageReferVenue=CoRL'22, Pipeline=Two-stage (gt)2025.02 | — | — | — | — | — | 43.9 |