Research map / Detection and segmentation / Open-set and few-shot detection
End-to-End Zero-Shot HOI Detection via Vision and Language Knowledge Distillation
Mingrui Wu, Jiaxin Gu, Yunhang Shen, Mingbao Lin, Chao Chen, Xiaoshuai Sun
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use EoID rather than GEN-VLKT | HOI detection · HICO-Det RF-UC | mAP (Unseen): 21.36 -> 22.04 | Improved |
| Use EoID rather than GEN-VLKT | HOI detection · HICO-Det NF-UC | mAP (Unseen): 25.05 -> 26.77 | Improved |
| Use EoID rather than GEN-VLKT | HOI detection · HICO-Det UC | mAP (Unseen): 20.64±0.89 -> 23.01±1.54 | Improved |
| Use EoID rather than GEN-VLKT | HOI detection · HICO-Det UA | mAP (Unseen): 20.85 -> 23.04 | Improved |
| Use EoID rather than GEN-VLKT | HOI detection · HICO-Det UV | mAP (Unseen): 20.96 -> 22.71 | Improved |
| Add unseen-action-only CLIP distillation (ablation) | EoID in HICO-Det UA · HICO-Det UA | mAP (Seen): 29.15 -> 30.20 | Improved |
| Add unseen-action-only CLIP distillation (ablation) | EoID in HICO-Det UA · HICO-Det UA | mAP (Unseen): 21.84 -> 21.98 | Improved |
| Detach gradients between the human-object decoder and interaction decoder (ablation) | EoID in HICO-Det UA · HICO-Det UA | mAP (Full): 27.93 -> 28.68 | Improved |
| Combine unseen-action-only distillation with decoder gradient detachment (ablation) | EoID in HICO-Det UA · HICO-Det UA | mAP (Full): 27.93 -> 29.22 | Improved |
| Increase topk across tested settings (ablation) | EoID · HICO-Det test, UA | mAP (Unseen): 21.96 -> 22.29 | Improved |
| Use topk 3 rather than topk 1 (ablation) | EoID · HICO-Det test, UA | mAP (Full): 28.45 -> 28.83 | Improved |
| Set interactive threshold to 0.5 (ablation) | EoID · HICO-Det test, UA | mAP (Full): 28.50 -> 28.83 | Improved |
Closest papers by meaning
- Exploring the Potential of Large Foundation Models for Open-Vocabulary HOI Detection (CVPR 2024 · similarity 0.91)
- Locality-Aware Zero-Shot Human-Object Interaction Detection (CVPR 2025 · similarity 0.91)
- Toward Open-Set Human Object Interaction Detection (AAAI 2024 · similarity 0.91)
- EZ-HOI: VLM Adaptation via Guided Prompt Learning for Zero-Shot HOI Detection (NeurIPS 2024 · similarity 0.91)
- Open-World Human-Object Interaction Detection via Multi-modal Prompts (CVPR 2024 · similarity 0.90)
- Visual Diversity and Region-aware Prompt Learning for Zero-shot HOI Detection (NeurIPS 2025 · similarity 0.90)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
