Persistence diagram reconstruction on ABC
0.4019W2PointGPT
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| PointGPTEncoder training protocol=Pretrained-frozen, Feature selection=Last transformer block (L)2026.04 | 0.4019 | 0.0317 | 3.799 | |
| PointGPTEncoder training protocol=Pretrained-frozen, Feature selection=Combined features from all transformer blocks (C)2026.04 | 0.4047 | 0.0321 | 3.806 | |
| Point-BERTEncoder training protocol=Pretrained-frozen, Feature selection=Combined features from all transformer blocks (C)2026.04 | 0.4323 | 0.032 | 3.78 | |
| Point-BERTEncoder training protocol=Pretrained-frozen, Feature selection=Last transformer block (L)2026.04 | 0.4337 | 0.0319 | 3.867 | |
| PCP-MAEEncoder training protocol=Pretrained-frozen, Feature selection=Last transformer block (L)2026.04 | 0.4509 | 0.0321 | 4.008 | |
| DGCNNEncoder training protocol=Encoders trained end-to-end2026.04 | 0.4521 | 0.0317 | 3.76 | |
| Point-MAEEncoder training protocol=Pretrained-frozen, Feature selection=Last transformer block (L)2026.04 | 0.4615 | 0.0321 | 3.868 | |
| PCP-MAEEncoder training protocol=Pretrained-frozen, Feature selection=Combined features from all transformer blocks (C)2026.04 | 0.4683 | 0.0321 | 3.793 | |
| Point-MAEEncoder training protocol=Pretrained-frozen, Feature selection=Combined features from all transformer blocks (C)2026.04 | 0.4799 | 0.0327 | 4.393 | |
| PointNetEncoder training protocol=Encoders trained end-to-end2026.04 | 0.5744 | 0.0323 | 3.849 | |
| PointNet++Encoder training protocol=Encoders trained end-to-end2026.04 | 0.9539 | 0.0321 | 39.13 |