Set containment on ModelNet40 |S|=128
98AccuracyMASNET-ReLU
Evaluation results
| Method | Links | |
|---|---|---|
| MASNET-ReLU|S|=128, Class ratio=1:1, Noise (std)=0.005, m (embedding dimension)=50, d (point dimension)=32025.10 | 98 | |
| MASNET-Hat|S|=128, Class ratio=1:1, Noise (std)=0.005, m (embedding dimension)=50, d (point dimension)=32025.10 | 87 | |
| FlexSubNet|S|=128, Class ratio=1:1, Noise (std)=0.005, m (embedding dimension)=50, d (point dimension)=32025.10 | 84 | |
| SetTransformer|S|=128, Class ratio=1:1, Noise (std)=0.005, m (embedding dimension)=50, d (point dimension)=32025.10 | 62 | |
| DeepSets|S|=128, Class ratio=1:1, Noise (std)=0.005, m (embedding dimension)=50, d (point dimension)=32025.10 | 52 | |
| Neural SFE|S|=128, Class ratio=1:1, Noise (std)=0.005, m (embedding dimension)=50, d (point dimension)=32025.10 | 52 |