Image-to-3D generation on Google Scanned Objects (GSO)
77.53CLIP SimilarityDehallu3D
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Dehallu3D2026.03 | 77.53 | 21.8407 | 0.0202 | — | — | 89.66 | 0.1453 | 42.12 | |
| InstantMesh2026.03 | 75.38 | 20.8954 | 0.022 | — | — | 89.03 | 0.1749 | 40.46 | |
| Unique3D2026.03 | 74.93 | 20.9795 | 0.0218 | — | — | 88.82 | 0.1742 | 40.73 | |
| SF3D2026.03 | 74.63 | 21.3257 | 0.0214 | — | — | 89.12 | 0.1537 | 37.65 | |
| CRM2026.03 | 71.91 | 21.1265 | 0.0216 | — | — | 88.89 | 0.172 | 39.67 | |
| TripoSR2026.03 | 71.46 | 20.5309 | 0.0224 | — | — | 88.97 | 0.1841 | 38.47 | |
| Wonder3D2026.03 | 69.7 | 20.4963 | 0.0218 | — | — | 89.08 | 0.1851 | 35.8 | |
| TriMMTraining Data Size=80K2025.08 | 52.5 | 14.34 | 0.034 | 60.7 | 78.6 | — | — | — | |
| TrellisTraining Data Size=500K2025.08 | 51.4 | 14.27 | 0.035 | 61.2 | 77.5 | — | — | — | |
| InstantMeshTraining Data Size=270K2025.08 | 50.5 | 14.17 | 0.058 | 45.7 | 65.4 | — | — | — | |
| LGMTraining Data Size=80K2025.08 | 50.3 | 14.32 | 0.072 | 32.7 | 52.5 | — | — | — | |
| CRMTraining Data Size=376K2025.08 | 49.7 | 13.83 | 0.043 | 56.2 | 75.6 | — | — | — | |
| BaselineDescription=Generative model trained only on RGB triplanes2025.08 | 48.2 | 13.37 | 0.061 | 42.7 | 59.1 | — | — | — | |
| TripoSR2025.08 | 47.8 | 13.05 | 0.096 | 40.9 | 56.4 | — | — | — |