Few-shot 3D Semantic Segmentation on ScanNet
68.86mIoUPENet
Evaluation Results
| Method | Links | |
|---|---|---|
| PENet#Params (Total)=25.09M, #Params (Trainable)=3.00M, FLOPs (G)=31.04, Setting=2-way 1-shot2025.09 | 68.86 | |
| PAP3D#Params (Total)=2.79M, #Params (Trainable)=2.79M, FLOPs (G)=16.30, Setting=2-way 1-shot2025.09 | 55.94 | |
| UPLN-way=1, K-shot=52026.03 | 46.83 | |
| CosegN-way=1, K-shot=52026.03 | 46.21 | |
| QGEN-way=1, K-shot=52026.03 | 43.49 | |
| UPLN-way=1, K-shot=12026.03 | 43 | |
| CosegN-way=1, K-shot=12026.03 | 41.78 | |
| AttMPTI#Params (Total)=357.82K, #Params (Trainable)=357.82K, FLOPs (G)=152.65, Setting=2-way 1-shot2025.09 | 41.69 | |
| QPGAN-way=1, K-shot=52026.03 | 39.94 | |
| UPLN-way=2, K-shot=52026.03 | 38.4 | |
| AttMPTIN-way=1, K-shot=52026.03 | 38.12 | |
| QGEN-way=1, K-shot=12026.03 | 35.2 | |
| CosegN-way=2, K-shot=52026.03 | 34.68 | |
| QPGAN-way=1, K-shot=12026.03 | 33.97 | |
| AttMPTIN-way=1, K-shot=12026.03 | 32.5 | |
| UPLN-way=2, K-shot=12026.03 | 32.39 | |
| QGEN-way=2, K-shot=52026.03 | 29.92 | |
| QPGAN-way=2, K-shot=52026.03 | 29.18 | |
| AttMPTIN-way=2, K-shot=52026.03 | 28.88 | |
| CosegN-way=2, K-shot=12026.03 | 28.78 | |
| QGEN-way=2, K-shot=12026.03 | 26.01 | |
| AttMPTIN-way=2, K-shot=12026.03 | 24.94 | |
| QPGAN-way=2, K-shot=12026.03 | 21.67 |