Video To Image Affordance Grounding On Opra
평가 지표
KLD
Top-1 Action Accuracy
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | KLD | Top-1 Action Accuracy | Paper Title | Repository |
|---|---|---|---|---|
| Afformer (ViTDet-B encoder) | 1.51 | 52.27 | Affordance Grounding from Demonstration Video to Target Image | |
| Afformer (ResNet-50-FPN encoder) | 1.55 | 52.14 | Affordance Grounding from Demonstration Video to Target Image | |
| Demo2Vec | 2.34 | 40.79 | Demo2Vec: Reasoning Object Affordances From Online Videos | - |
0 of 3 row(s) selected.