3D Classification on ScanObjectNN
67Top-1 Accuracy3DAlign-DAER
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| 3DAlign-DAERZero-shot=true2025.11 | 67 | 86.2 | 93.1 | |
| SAGE-7BTraining Dataset=Objaverse paired with Cap3D generated captions, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 65.8 | 80.2 | 90.6 | |
| Uni3D-gVenue=ICLR 2024, Zero-shot=true2025.11 | 65.2 | 82.9 | 91.3 | |
| Uni3D-LVenue=ICLR 2024, Zero-shot=true2025.11 | 64.1 | 81.9 | 90.5 | |
| Uni3D-BTraining Dataset=Ensembled, Teacher Backbone=EVA-CLIP-E [52]2026.03 | 63.8 | 82.7 | 90.2 | |
| Uni3D-BVenue=ICLR 2024, Zero-shot=true2025.11 | 63.7 | 81.7 | 90.3 | |
| RECON++-BTraining Dataset=Ensembled, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 63.6 | 84.2 | 90.6 | |
| ReCon++-L (shapeLLM)Venue=ECCV 2024, Zero-shot=true2025.11 | 63.5 | 80.1 | 90.5 | |
| MixCon3DTraining Dataset=Ensembled, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 58.6 | 80.3 | 89.2 | |
| Uni3D-LTraining Dataset=Ensembled, Teacher Backbone=EVA-CLIP-E [52]2026.03 | 58.2 | 81.8 | 89.4 | |
| TAMMTraining Dataset=ShapeNet, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 54.8 | 74.5 | 83.3 | |
| MixCon3DTraining Dataset=ShapeNet, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 52.6 | 69.9 | 78.7 | |
| OpenShapeTraining Dataset=Ensembled, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 52.2 | 79.7 | 88.7 | |
| OpenShape-PointBERTVenue=NeurIPS 2023, Zero-shot=true2025.11 | 52.1 | 79.6 | 88.8 | |
| ULIP-2Venue=CVPR 2024, Zero-shot=true2025.11 | 51.7 | 72.6 | 82.4 | |
| ULIP-2Training Dataset=Ensembled, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 51.6 | 72.5 | 82.3 | |
| ULIPTraining Dataset=ShapeNet, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 51.5 | 71.1 | 80.2 | |
| ULIPVenue=CVPR 2023, Zero-shot=true2025.11 | 51.4 | 71 | 80.3 | |
| OpenShapeTraining Dataset=ShapeNet, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 47.2 | 72.4 | 84.7 | |
| RECONTraining Dataset=ShapeNet, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 42.3 | 62.5 | 75.6 | |
| PointCLIPv23D Training status=2D Inference without 3D Training, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 42.2 | 63.3 | 74.5 | |
| CLIP2PointTraining Dataset=ShapeNet, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 25.5 | 44.6 | 59.4 | |
| PointCLIP3D Training status=2D Inference without 3D Training, Teacher Backbone=OpenCLIP-bigG [30]2026.03 | 10.5 | 20.8 | 30.6 | |
| PointCLIPVenue=CVPR 2022, Zero-shot=true2025.11 | 10.4 | 20.9 | 30.5 |