Shape Reconstruction on ShapeNet Sparse Scans
0.0088Cars CDGG-Langevin
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| GG-LangevinMethod category=Learned prior2026.03 | 0.0088 | 25.4 | 0.0063 | 17.7 | 0.0122 | 14.3 | 0.0104 | 17 | |
| IGRMethod category=Optimization2026.03 | 0.0107 | 27.7 | 0.028 | 33.4 | 0.0236 | 25.1 | 0.0252 | 29.1 | |
| NKSRMethod category=Learned prior2026.03 | 0.0117 | 28.6 | 0.0126 | 29.7 | 0.0144 | 18.6 | 0.0131 | 21.5 | |
| DiffCDMethod category=Optimization2026.03 | 0.0122 | 29.3 | 0.0088 | 25.7 | 0.0148 | 21 | 0.0158 | 24.1 | |
| DeepSDFMethod category=Learned prior2026.03 | 0.0126 | 34.4 | 0.0155 | 39.5 | 0.0151 | 25.8 | 0.0165 | 30.4 | |
| 3DILGMethod category=Learning-based2026.03 | 0.0168 | 30.6 | 0.0101 | 25.9 | 0.0137 | 23.1 | 0.0138 | 26.7 | |
| VecSetMethod category=Learning-based2026.03 | 0.0181 | 32.5 | 0.0099 | 26.8 | 0.0129 | 17.7 | 0.0123 | 22.6 | |
| ShapeFormerMethod category=Learning-based2026.03 | 0.0192 | 32 | 0.0218 | 33.5 | 0.021 | 20.2 | 0.0266 | 28.6 |