Human-Object Interaction Detection on HICO-DET Non-rare First Unseen Composition (NF-UC)
36.33AP (Unseen)EZ-HOI
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| EZ-HOIBackbone=Resnet50+ViT-L, Setting=NF2024.10 | 36.33 | 34.47 | 34.84 | |
| EZ-HOIBackbone=Resnet50+ViT-L, Setting=NF2024.10 | 36.33 | 34.47 | 34.84 | |
| GRASP-HOISetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 35.61 | 38.95 | 38.24 | |
| SICHOISetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 34.52 | 36.06 | 35.75 | |
| HOIGenSetting=NF-UC2024.08 | 33.98 | 32.86 | 33.08 | |
| HOIGenSetting=NF-UC2024.08 | 33.98 | 32.86 | 33.08 | |
| HOIGenSetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 33.98 | 32.86 | 33.08 | |
| EZ-HOIBackbone=Resnet50+ViT-B, Setting=NF2024.10 | 33.66 | 30.55 | 31.17 | |
| EZ-HOIBackbone=Resnet50+ViT-B, Setting=NF2024.10 | 33.66 | 30.55 | 31.17 | |
| CMMPSetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 33.52 | 35.53 | 35.13 | |
| BCOMSetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 33.12 | 31.76 | 32.03 | |
| BC-HOISetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 33.01 | 37.24 | 36.4 | |
| ADA-CMSetting=NF-UC2024.08 | 32.41 | 31.13 | 31.39 | |
| ADA-CMSetting=NF-UC2024.08 | 32.41 | 31.13 | 31.39 | |
| ADA-CMBackbone=Resnet50+ViT-B, Setting=NF2024.10 | 32.41 | 31.13 | 31.39 | |
| ADA-CMBackbone=Resnet50+ViT-B, Setting=NF2024.10 | 32.41 | 31.13 | 31.39 | |
| CLIP4HOISetting=NF-UC2024.08 | 31.44 | 28.26 | 28.9 | |
| CLIP4HOISetting=NF-UC2024.08 | 31.44 | 28.26 | 28.9 | |
| CLIP4HOIBackbone=Resnet50+ViT-B, Setting=NF2024.10 | 31.44 | 28.26 | 28.9 | |
| CLIP4HOIBackbone=Resnet50+ViT-B, Setting=NF2024.10 | 31.44 | 28.26 | 28.9 | |
| DP-HOISetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 28.87 | 29.98 | 29.76 | |
| UniHOI-s (w/ BLIP2)Type=NF-UC, Zero-shot=true2023.11 | 28.45 | 32.63 | 31.79 | |
| UniHOIBackbone=Resnet50+ViT-L, Setting=NF2024.10 | 28.45 | 32.63 | 31.79 | |
| UniHOIBackbone=Resnet50+ViT-L, Setting=NF2024.10 | 28.45 | 32.63 | 31.79 | |
| UniHOI (BLIP2)Setting=Zero-shot, Protocol=Open-vocabulary2025.12 | 28.45 | 32.63 | 31.79 | |
| LOGICHOISetting=NF-UC2024.08 | 26.84 | 27.86 | 27.95 | |
| LOGICHOISetting=NF-UC2024.08 | 26.84 | 27.86 | 27.95 | |
| EoIDBackbone=Resnet50, Setting=NF2024.10 | 26.77 | 26.66 | 26.69 | |
| EoIDBackbone=Resnet50, Setting=NF2024.10 | 26.77 | 26.66 | 26.69 | |
| FleVRSevaluation_mode=zero-shot2024.08 | 26.62 | 31.17 | 30.17 | |
| HOICLIPSetting=NF-UC2024.08 | 26.39 | 28.1 | 27.75 | |
| HOICLIPSetting=NF-UC2024.08 | 26.39 | 28.1 | 27.75 | |
| HOICLIPBackbone=Resnet50+ViT-B, Setting=NF2024.10 | 26.39 | 28.1 | 27.75 | |
| HOICLIPBackbone=Resnet50+ViT-B, Setting=NF2024.10 | 26.39 | 28.1 | 27.75 | |
| HOICLIPSetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 26.39 | 28.1 | 27.75 | |
| GEN-VLKTType=NF-UC, Zero-shot=true2023.11 | 25.05 | 23.38 | 23.71 | |
| GEN-VLKTSetting=NF-UC2024.08 | 25.05 | 23.38 | 23.71 | |
| GEN-VLKTSetting=NF-UC2024.08 | 25.05 | 23.38 | 23.71 | |
| GEN-VLKTevaluation_mode=zero-shot2024.08 | 25.05 | 23.38 | 23.71 | |
| GEN-VLKTBackbone=Resnet50+ViT-B, Setting=NF2024.10 | 25.05 | 23.38 | 23.71 | |
| GEN-VLKTBackbone=Resnet50+ViT-B, Setting=NF2024.10 | 25.05 | 23.38 | 23.71 | |
| GEN-VLKTSetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 25.05 | 23.38 | 23.71 | |
| RLIPv2-ParSeDASetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 22.81 | 29.52 | 28.18 | |
| FCLType=NF-UC, Zero-shot=true2023.11 | 18.66 | 19.55 | 19.37 | |
| FCLevaluation_mode=zero-shot2024.08 | 18.66 | 19.55 | 19.37 | |
| ATLType=NF-UC, Zero-shot=true2023.11 | 18.25 | 18.78 | 18.67 | |
| ATLevaluation_mode=zero-shot2024.08 | 18.25 | 18.78 | 18.67 | |
| VCLType=NF-UC, Zero-shot=true2023.11 | 16.22 | 18.52 | 18.06 | |
| VCLevaluation_mode=zero-shot2024.08 | 16.22 | 18.52 | 18.06 |