Human-Object Interaction Recognition on HICO (Systematic-hard)
23.06mAP (Full Classes)RelViT + EsViT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| RelViT + EsViTExtra Supervision=None, Backbone=PVTv2-b22022.04 | 23.06 | 22.89 | |
| RelViTExtra Supervision=None, Backbone=PVTv2-b22022.04 | 22.75 | 22.66 | |
| EsViTExtra Supervision=None, Backbone=PVTv2-b22022.04 | 22.55 | 21.84 | |
| Hou et al.Extra Supervision=None, Backbone=ResNet-1012022.04 | 21.76 | 10.58 | |
| ViT-onlyExtra Supervision=None, Backbone=PVTv2-b22022.04 | 19.03 | 18.85 |