Human-Object Interaction Detection on V-COCO (AP Role Metrics)
73.4AP^1 RoleHOI-IDiff
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| HOI-IDiff2025.03 | 73.4 | 76.1 | — | — | |
| GRASP-HOIBackbone=ResNet-50+MLLM, Closed-set setting=true2025.12 | 72.5 | 76.2 | — | — | |
| Yuan et al.2025.03 | 72.1 | 74.1 | — | — | |
| RLIPv2Backbone=Swin-L, Closed-set setting=true2025.12 | 72.1 | 74.1 | — | — | |
| Luo et al.2025.03 | 71.1 | 75.5 | — | — | |
| SICHOIBackbone=R101+VIT-L, Closed-set setting=true2025.12 | 71.1 | 75.6 | — | — | |
| HORPBackbone=Swin-L+CLIP, Closed-set setting=true2025.12 | 68.9 | 71.1 | — | — | |
| BC-HOIBackbone=ResNet-50, Closed-set setting=true2025.12 | 68.2 | 70.1 | — | — | |
| UniHOI-1Backbone=R101+VIT-L, Closed-set setting=true2025.12 | 68.1 | 70.8 | — | — | |
| DIFFUSIONHOIBackbone=ViT-L, VL Pretrain=Stable unCLIP2024.10 | 67.1 | 71.1 | — | — | |
| DIFFUSIONHOIBackbone=VQGAN, VL Pretrain=Stable Diffusion2024.10 | 66.8 | 70.9 | — | — | |
| DP-HOIBackbone=ResNet-50, Closed-set setting=true2025.12 | 66.6 | — | — | — | |
| CQLBackbone=ResNet-101, Closed-set setting=true2025.12 | 66.5 | 69.9 | — | — | |
| CQLBackbone=R50, VL Pretrain=CLIP2024.10 | 66.4 | 69.2 | — | — | |
| Yang et al.2025.03 | 66.2 | 67.6 | — | — | |
| MP-HOIBackbone=Swin-L+ViT, Closed-set setting=true2025.12 | 66.2 | 67.6 | — | — | |
| STIPBackbone=R50, VL Pretrain=-2024.10 | 66 | 70.7 | — | — | |
| Wang et al.2025.03 | 65.8 | 69.9 | — | — | |
| BCOMBackbone=ResNet-50, Closed-set setting=true2025.12 | 65.8 | 69.9 | — | — | |
| AGERBackbone=R50, VL Pretrain=CLIP2024.10 | 65.7 | 69.7 | — | — | |
| AGERBackbone=ResNet-50, Closed-set setting=true2025.12 | 65.7 | 69.7 | — | — | |
| DiffHOIBackbone=Swin-L, Closed-set setting=true2025.12 | 65.7 | 68.2 | — | — | |
| RmLRBackbone=ResNet-101, Closed-set setting=true2025.12 | 64.2 | 70.2 | — | — | |
| PVICBackbone=Swin-L, VL Pretrain=-2024.10 | 64.1 | 70.2 | — | — | |
| Zhang et al. [63]2025.03 | 64.1 | 70.2 | — | — | |
| PVICBackbone=Swin-L, Closed-set setting=true2025.12 | 64.1 | 70.2 | — | — | |
| RmLRBackbone=R50, VL Pretrain=MobileBERT2024.10 | 63.8 | 69.8 | — | — | |
| SSRTBackbone=R50, VL Pretrain=CLIP2024.10 | 63.7 | 65.9 | — | — | |
| Liao et al.2025.03 | 63.6 | 65.9 | — | — | |
| DOQBackbone=R50, VL Pretrain=CLIP2024.10 | 63.5 | — | — | — | |
| HOICLIPBackbone=R50, VL Pretrain=CLIP2024.10 | 63.5 | 64.8 | — | — | |
| HOICLIPBackbone=ResNet-50, Closed-set setting=true2025.12 | 63.5 | 64.8 | — | — | |
| CPChoiBackbone=R50, VL Pretrain=-2024.10 | 63.1 | 65.4 | — | — | |
| IF-HOIBackbone=R50, VL Pretrain=-2024.10 | 63 | 65.2 | — | — | |
| MCPCBackbone=R50, VL Pretrain=-2024.10 | 63 | 65.1 | — | — | |
| Wu et al.2025.03 | 63 | 68.7 | — | — | |
| PAFRBackbone=Swin-L, Closed-set setting=true2025.12 | 63 | 68.7 | — | — | |
| PVICBackbone=R50, VL Pretrain=-2024.10 | 62.8 | 67.8 | — | — | |
| GEN-VLKBackbone=R50, VL Pretrain=CLIP2024.10 | 62.4 | 64.4 | — | — | |
| GEN-VLKTBackbone=R50+CLIP, Closed-set setting=true2025.12 | 62.4 | 64.7 | — | — | |
| CDNBackbone=R50, VL Pretrain=-2024.10 | 62.3 | 64.4 | — | — | |
| CDNBackbone=ResNet-50, Closed-set setting=true2025.12 | 62.3 | 64.4 | — | — | |
| VIPLOBackbone=ViT-B, VL Pretrain=CLIP2024.10 | 62.2 | 68 | — | — | |
| VIPLOBackbone=R50+CLIP, Closed-set setting=true2025.12 | 62.2 | 68 | — | — | |
| MSTRBackbone=R50, VL Pretrain=-2024.10 | 62 | 65.2 | — | — | |
| RLIPBackbone=ResNet-50, Closed-set setting=true2025.12 | 61.9 | 64.2 | — | — | |
| Zhang et al. [62]2025.03 | 61.3 | 67.1 | — | — | |
| UPTBackbone=ResNet-101, Closed-set setting=true2025.12 | 61.3 | 67.1 | — | — | |
| IwinBackbone=R50-FPN, VL Pretrain=-2024.10 | 60.5 | — | — | — | |
| FGAHOIBackbone=Swin-L, Closed-set setting=true2025.12 | 60.5 | 61.2 | — | — | |
| EZ-HOIBackbone=ResNet-50, Closed-set setting=true2025.12 | 60.5 | 66.2 | — | — | |
| CTANBackbone=R50, VL Pretrain=CLIP2024.10 | 60.1 | — | — | — | |
| UPTBackbone=R50, VL Pretrain=-2024.10 | 59 | 64.5 | — | — | |
| QPICBackbone=R50, VL Pretrain=-2024.10 | 58.8 | 61 | — | — | |
| Tamura et al.2025.03 | 58.8 | 61 | — | — | |
| QPICBackbone=ResNet-101, Closed-set setting=true2025.12 | 58.8 | 61 | — | — | |
| ADA-CMBackbone=ViT-L, VL Pretrain=CLIP2024.10 | 58.6 | 64 | — | — | |
| ADA-CMBackbone=R50+CLIP, Closed-set setting=true2025.12 | 58.6 | 64 | — | — | |
| HOTRBackbone=R50, VL Pretrain=-2024.10 | 55.2 | 64.4 | — | — | |
| AS-NetBackbone=R50, VL Pretrain=-2024.10 | 53.9 | — | — | — | |
| DRGBackbone=R50-FPN, VL Pretrain=-2024.10 | 51 | — | — | — | |
| UnionDetBackbone=R50, VL Pretrain=-2024.10 | 47.5 | 56.2 | — | — | |
| Kim et al.2025.03 | 47.5 | 56.2 | — | — | |
| iCANBackbone=R50, VL Pretrain=-2024.10 | 45.3 | — | — | — | |
| Gkioxari et al.2025.03 | 40 | 48 | — | — | |
| ADA-CMSupervision Setting=fully supervised2026.02 | — | — | — | 61.2 | |
| BC-HOIFeature Extractor=BLIP2, Supervision Setting=fully supervised2026.02 | — | — | 68.2 | 70.6 | |
| BCOMSupervision Setting=fully supervised2026.02 | — | — | 65.8 | 69.9 | |
| CLIP4HOIBackbone=ResNet-50, Closed-set setting=true2025.12 | — | 66.3 | — | — | |
| CLIP4HOISupervision Setting=fully supervised2026.02 | — | — | — | 66.3 | |
| CMMPBackbone=ResNet-50, Closed-set setting=true2025.12 | — | 64 | — | — | |
| CMMPSupervision Setting=fully supervised2026.02 | — | — | — | 64 | |
| DA-HOISupervision Setting=fully supervised2026.02 | — | — | 69.4 | 71.7 | |
| DA-HOIDetector=Grounding-DINO, Supervision Setting=fully supervised2026.02 | — | — | 70.4 | 72.5 | |
| DA-HOIDetector=Yolo-World, Supervision Setting=fully supervised2026.02 | — | — | 70.1 | 72.4 | |
| EZ-HOISupervision Setting=fully supervised2026.02 | — | — | 60.5 | 66.2 | |
| GEN-VLKTSupervision Setting=fully supervised2026.02 | — | — | 62.4 | 64.5 | |
| HOICLIPSupervision Setting=fully supervised2026.02 | — | — | 63.5 | 64.8 | |
| LAINSupervision Setting=fully supervised2026.02 | — | — | — | 65.1 | |
| UniHOIFeature Extractor=BLIP2, Supervision Setting=fully supervised2026.02 | — | — | 65.6 | 68.3 |