List your company

Human-Object Interaction Detection on HICO-DET Unseen Verb (UV)

28.82Unseen ScoreEZ-HOI
8.415213.712619.0124.3074Nov 7, 2023Apr 6, 2024Sep 5, 2024Feb 4, 2025Jul 6, 2025Dec 5, 2025May 6, 2026

Evaluation results

MethodLinks
EZ-HOIBackbone=Resnet50+ViT-L, Trainable Parameters=14.07M, Setting=UV2024.1028.8238.1536.84—
ScriptHOISetting=Open-vocabulary2026.052834—30.7
UniHOI-s (w/ BLIP2)Type=UV, Zero-shot=true2023.1126.0536.7834.68—
UniHOIBackbone=Resnet50+ViT-L, Trainable Parameters=52.3M, Setting=UV2024.1026.0536.7834.68—
CLIP4HOIBackbone=Resnet50+ViT-B, Trainable Parameters=56.7M, Setting=UV2024.1026.0231.1430.42—
EZ-HOIBackbone=Resnet50+ViT-B, Trainable Parameters=6.85M, Setting=UV2024.1025.133.4932.32—
HOICLIPBackbone=Resnet50+ViT-B, Trainable Parameters=66.18M, Setting=UV2024.1024.332.1931.09—
Loc-aware OV-HOISetting=Open-vocabulary2026.0524.333.1—28
EoIDBackbone=Resnet50, Trainable Parameters=41.45M, Setting=UV2024.1022.7130.7329.61—
CLIP4HOISetting=Open-vocabulary2026.0522.532.6—26.6
FleVRSevaluation_mode=zero-shot2024.0821.535.6333.09—
GEN-VLKTType=UV, Zero-shot=true2023.1120.9630.2328.74—
GEN-VLKTevaluation_mode=zero-shot2024.0820.9630.2328.74—
GEN-VLKTBackbone=Resnet50+ViT-B, Trainable Parameters=42.05M, Setting=UV2024.1020.9630.2328.74—
GEN-VLKTSetting=Open-vocabulary2026.0518.431.8—23.3
QPIC-OVSetting=Open-vocabulary2026.0512.827.6—17.5
CLIP-ZSSetting=Open-vocabulary2026.059.223.5—13.2