3D-Text Matching on Objaverse-LVIS 1.0 (test)
67.6CLIP Matching AccuracyAffine + Subspace Projection
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Affine + Subspace Projection3D Encoder=OpenShape, 3D Encoder Category=Multi-modal2025.03 | 67.6 | 55.2 | 45 | |
| Affine + Subspace Projection3D Encoder=ULIP-2, 3D Encoder Category=Multi-modal2025.03 | 65.6 | 47.2 | 35.2 | |
| Affine + Subspace Projection3D Encoder=Uni3D, 3D Encoder Category=Multi-modal2025.03 | 61.4 | 45.8 | 34.4 | |
| Affine + Subspace Projection3D Encoder=PointBert, 3D Encoder Category=Uni-modal2025.03 | 30.8 | 23.2 | 15.6 | |
| Local CKA + Subspace Projection3D Encoder=PointBert, 3D Encoder Category=Uni-modal2025.03 | 29.4 | 17 | 15 | |
| Local CKA + Subspace Projection3D Encoder=Pointnet, 3D Encoder Category=Uni-modal2025.03 | 26.8 | 18.6 | 14.3 | |
| Affine + Subspace Projection3D Encoder=Pointnet, 3D Encoder Category=Uni-modal2025.03 | 25.2 | 22.4 | 16.6 | |
| Affine + Subspace Projection3D Encoder=SparseConv, 3D Encoder Category=Uni-modal2025.03 | 21.4 | 19.2 | 15.8 | |
| Local CKA + Subspace Projection3D Encoder=SparseConv, 3D Encoder Category=Uni-modal2025.03 | 19 | 15.8 | 10 | |
| Affine3D Encoder=Pointnet, 3D Encoder Category=Uni-modal2025.03 | 18.4 | 8 | 9.6 | |
| Affine3D Encoder=PointBert, 3D Encoder Category=Uni-modal2025.03 | 15.8 | 7.8 | 6.4 | |
| Affine3D Encoder=SparseConv, 3D Encoder Category=Uni-modal2025.03 | 11 | 6 | 4.2 | |
| Local CKA3D Encoder=Pointnet, 3D Encoder Category=Uni-modal2025.03 | 6.6 | 2.4 | 1 | |
| Local CKA3D Encoder=PointBert, 3D Encoder Category=Uni-modal2025.03 | 5.8 | 1.8 | 1 | |
| Local CKA3D Encoder=SparseConv, 3D Encoder Category=Uni-modal2025.03 | 3.4 | 1.79 | 0.6 |