Classification on ModelNet40 GZSL v1 (seen + unseen test)
84.7Accuracy (Seen)CADA-VAE*
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CADA-VAE*Backbone=PointNet, Embedding=W2V, Generative=true2021.08 | 84.7 | 1.3 | 2.6 | |
| f-CLSWGAN*Backbone=PointNet, Embedding=W2V, Generative=true2021.08 | 76.3 | 3.7 | 7 | |
| MHPCBackbone=PointNet, Embedding=W2V, Bias reduct.=true2021.08 | 53.8 | 26.2 | 35.2 | |
| MHPCBackbone=PointNet, Embedding=GloVe, Bias reduct.=true2021.08 | 53.8 | 25.7 | 34.8 | |
| ZSLPCBackbone=PointNet + NetVlad, Embedding=GloVe2021.08 | 49.2 | 18.2 | 26.6 | |
| 3DGenZBackbone=PointNet, Embedding=W2V, Generative=true, Bias reduct.=true2021.08 | 48.8 | 29.3 | 36.6 | |
| 3DGenZBackbone=PointNet, Embedding=GloVe + W2V, Generative=true, Bias reduct.=true2021.08 | 47.8 | 36.5 | 41.3 | |
| 3DGenZBackbone=PointNet, Embedding=GloVe, Generative=true, Bias reduct.=true2021.08 | 44.7 | 28.4 | 34.7 | |
| ZSLPCBackbone=PointNet + NetVlad, Embedding=W2V2021.08 | 40.1 | 22.5 | 28.8 |