Human-Object Interaction Detection on HICO-DET (Rare First Unseen Combination (RF-UC))
44.81mAP (Full)DA-HOI
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DA-HOIDetector=Grounding-DINO2026.02 | 44.81 | 45.19 | 43.3 | |
| DA-HOIDetector=Yolo-World2026.02 | 44 | 44.05 | 43.81 | |
| DA-HOIDetector=ResNet50 DETR2026.02 | 43.56 | 44.01 | 41.79 | |
| GRASP-HOISetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 42.46 | 42.98 | 43.67 | |
| BC-HOISetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 40.99 | 40.67 | 42.31 | |
| BC-HOIFeatures=BLIP22026.02 | 40.99 | 40.67 | 42.31 | |
| SICHOISetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 40.11 | 41.58 | 34.24 | |
| LAINBackbone=ViT-L2025.05 | 38.13 | 38.54 | 36.57 | |
| CMMPSetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 37.13 | 37.42 | 35.98 | |
| EZ-HOIBackbone=Resnet50+ViT-L, Setting=RF2024.10 | 36.73 | 37.35 | 34.24 | |
| EZ-HOIBackbone=Resnet50+ViT-L, Setting=RF2024.10 | 36.73 | 37.35 | 34.24 | |
| EZ-HOISetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 36.73 | 37.35 | 34.24 | |
| EZ-HOI2026.02 | 36.73 | 37.35 | 34.24 | |
| FleVRSevaluation_mode=zero-shot2024.08 | 36.6 | 39.61 | 26.06 | |
| DiffHOISetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 35.89 | 36.77 | 32.06 | |
| DP-HOISetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 35.03 | 36.17 | 30.49 | |
| LAINBackbone=ViT-B2025.05 | 34.41 | 35.06 | 31.83 | |
| LAIN2026.02 | 34.41 | 35.06 | 31.83 | |
| CLIP4HOISetting=RF-UC2024.08 | 34.08 | 35.48 | 28.47 | |
| CLIP4HOISetting=RF-UC2024.08 | 34.08 | 35.48 | 28.47 | |
| CLIP4HOIBackbone=Resnet50+ViT-B, Setting=RF2024.10 | 34.08 | 35.48 | 28.47 | |
| CLIP4HOIBackbone=Resnet50+ViT-B, Setting=RF2024.10 | 34.08 | 35.48 | 28.47 | |
| CLIP4HOIBackbone=ViT-B2025.05 | 34.08 | 35.48 | 28.47 | |
| CLIP4HOI2026.02 | 34.08 | 35.48 | 28.47 | |
| HOIGenSetting=RF-UC2024.08 | 33.86 | 34.57 | 31.01 | |
| HOIGenSetting=RF-UC2024.08 | 33.86 | 34.57 | 31.01 | |
| BCOMBackbone=ViT-L2025.05 | 33.74 | 35.04 | 28.52 | |
| BCOMDetector=ResNet50 DETR2026.02 | 33.74 | 35.04 | 28.52 | |
| LOGICHOISetting=RF-UC2024.08 | 33.17 | 34.93 | 25.97 | |
| LOGICHOISetting=RF-UC2024.08 | 33.17 | 34.93 | 25.97 | |
| LOGICHOIBackbone=ViT-B2025.05 | 33.17 | 34.93 | 25.97 | |
| EZ-HOIBackbone=Resnet50+ViT-B, Setting=RF2024.10 | 33.13 | 34.15 | 29.02 | |
| EZ-HOIBackbone=Resnet50+ViT-B, Setting=RF2024.10 | 33.13 | 34.15 | 29.02 | |
| ADA-CMSetting=RF-UC2024.08 | 33.01 | 34.35 | 27.63 | |
| ADA-CMSetting=RF-UC2024.08 | 33.01 | 34.35 | 27.63 | |
| ADA-CMBackbone=Resnet50+ViT-B, Setting=RF2024.10 | 33.01 | 34.35 | 27.63 | |
| ADA-CMBackbone=Resnet50+ViT-B, Setting=RF2024.10 | 33.01 | 34.35 | 27.63 | |
| ADA-CMBackbone=ViT-B2025.05 | 33.01 | 34.35 | 27.63 | |
| ADA-CMDetector=ResNet50 DETR2026.02 | 33.01 | 34.35 | 27.63 | |
| HOICLIPSetting=RF-UC2024.08 | 32.99 | 34.85 | 25.53 | |
| HOICLIPSetting=RF-UC2024.08 | 32.99 | 34.85 | 25.53 | |
| HOICLIPBackbone=Resnet50+ViT-B, Setting=RF2024.10 | 32.99 | 28.47 | 25.53 | |
| HOICLIPBackbone=Resnet50+ViT-B, Setting=RF2024.10 | 32.99 | 28.47 | 25.53 | |
| HOICLIPBackbone=ViT-B2025.05 | 32.99 | 34.85 | 25.53 | |
| HOICLIPSetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 32.99 | 34.85 | 25.53 | |
| HOICLIP2026.02 | 32.99 | 34.85 | 25.53 | |
| RLIPv2-ParSeDASetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 32.97 | 35.85 | 21.45 | |
| UniHOIBackbone=Resnet50+ViT-L, Setting=RF2024.10 | 32.27 | 33.16 | 28.68 | |
| UniHOIBackbone=Resnet50+ViT-L, Setting=RF2024.10 | 32.27 | 33.16 | 28.68 | |
| UniHOI (BLIP2)Setting=Zero-shot, Protocol=Open-vocabulary2025.12 | 32.27 | 33.16 | 28.68 | |
| UniHOIFeatures=BLIP22026.02 | 32.27 | 33.16 | 28.68 | |
| CMMPBackbone=ViT-B2025.05 | 32.18 | 32.87 | 29.45 | |
| CMMP2026.02 | 32.18 | 32.87 | 29.45 | |
| DA-HOITraining-free=true2026.02 | 31.5 | — | — | |
| GEN-VLKTSetting=RF-UC2024.08 | 30.56 | 32.91 | 21.36 | |
| GEN-VLKTSetting=RF-UC2024.08 | 30.56 | 32.91 | 21.36 | |
| GEN-VLKTevaluation_mode=zero-shot2024.08 | 30.56 | 32.91 | 21.36 | |
| GEN-VLKTBackbone=Resnet50+ViT-B, Setting=RF2024.10 | 30.56 | 32.91 | 21.36 | |
| GEN-VLKTBackbone=Resnet50+ViT-B, Setting=RF2024.10 | 30.56 | 32.91 | 21.36 | |
| GEN-VLKTBackbone=ViT-B2025.05 | 30.56 | 32.91 | 21.36 | |
| GEN-VLKTSetting=Zero-shot, Protocol=Open-vocabulary2025.12 | 30.56 | 32.91 | 21.36 | |
| GEN-VLKT2026.02 | 30.56 | 32.91 | 21.36 | |
| RLIPBackbone=ViT-B2025.05 | 30.52 | 33.35 | 19.19 | |
| EoIDBackbone=Resnet50, Setting=RF2024.10 | 29.52 | 31.39 | 22.04 | |
| EoIDBackbone=Resnet50, Setting=RF2024.10 | 29.52 | 31.39 | 22.04 | |
| EoIDBackbone=ViT-B2025.05 | 29.52 | 31.39 | 22.04 | |
| ADA-CMDetector=Grounding-DINO2026.02 | 28.26 | 30.12 | 20.82 | |
| ADA-CMDetector=Yolo-World2026.02 | 27.47 | 29.23 | 20.43 | |
| ADA-CMTraining-free=true2026.02 | 25.19 | — | — | |
| CLIPBackbone=ViT-B2025.05 | 23.36 | 22 | 28.79 | |
| FCLevaluation_mode=zero-shot2024.08 | 22.01 | 24.23 | 13.16 | |
| FCLBackbone=ViT-B2025.05 | 22.01 | 24.23 | 13.16 | |
| ATLevaluation_mode=zero-shot2024.08 | 21.57 | 24.67 | 9.18 | |
| ATLBackbone=ViT-B2025.05 | 21.57 | 24.67 | 9.18 | |
| VCLevaluation_mode=zero-shot2024.08 | 21.43 | 24.28 | 10.06 | |
| BCOMDetector=Grounding-DINO2026.02 | 20.31 | 23.08 | 9.23 | |
| BCOMDetector=Yolo-World2026.02 | 17.69 | 20.07 | 8.17 |