Research map / Detection and segmentation / Open-set and few-shot detection
Open-Vocabulary HOI Detection with Interaction-aware Prompt and Concept Calibration
Ting Lei, Shaofeng Yin, Qingchao Chen, Yuxin Peng, Yang Liu
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 |
|---|---|---|---|
| add interaction-aware prompts (INP) (ablation) | INP-CC · SWIG-HOI Full | mAP: 14.43 -> 15.54 | Improved |
| add concept calibration (CC) (ablation) | INP-CC · SWIG-HOI Full | mAP: 14.43 -> 15.30 | Improved |
| add CC on top of INP (ablation) | INP-CC · SWIG-HOI Full | mAP: 15.54 -> 16.74 | Improved |
| add INP on top of CC (ablation) | INP-CC · SWIG-HOI Full | mAP: 15.30 -> 16.74 | Improved |
| add CC on top of INP (ablation) | INP-CC · SWIG-HOI Unseen | mAP: 10.15 -> 11.02 | Improved |
| add interaction-aware prompts of size 128 without selection to common prompt (ablation) | INP-CC · SWIG-HOI Full | mAP: 15.30 -> 15.98 | Improved |
| add selection mechanism to prompts (ablation) | INP-CC · SWIG-HOI Full | mAP: 15.98 -> 16.74 | Improved |
| reduce prompt size from 128 to 64 (ablation) | INP-CC · SWIG-HOI Full | mAP: 16.74 -> 16.31 | Hurt |
| increase prompt size from 128 to 256 (ablation) | INP-CC · SWIG-HOI Full | mAP: 16.74 -> 16.18 | Hurt |
| cluster CLIP embeddings of descriptions instead of HOI names (ablation) | INP-CC · SWIG-HOI Full | mAP: 16.17 -> 15.93 | Hurt |
| cluster with Instructor T5 embeddings of descriptions instead of CLIP names (ablation) | INP-CC · SWIG-HOI Full | mAP: 16.17 -> 16.74 | Improved |
| use hard negative sampling instead of easy (ablation) | INP-CC · SWIG-HOI Full | mAP: 15.95 -> 16.74 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- CLIP4HOI — CLIP4HOI: Towards Adapting CLIP for Practical Zero-Shot HOI Detection (NeurIPS 2023) · better on 0 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Streamlined Open-Vocabulary Human-Object Interaction Detection (CVPR 2026 · similarity 0.95)
- Exploring the Potential of Large Foundation Models for Open-Vocabulary HOI Detection (CVPR 2024 · similarity 0.94)
- Open-World Human-Object Interaction Detection via Multi-modal Prompts (CVPR 2024 · similarity 0.94)
- Visual Diversity and Region-aware Prompt Learning for Zero-shot HOI Detection (NeurIPS 2025 · similarity 0.91)
- SGC-Net: Stratified Granular Comparison Network for Open-Vocabulary HOI Detection (CVPR 2025 · similarity 0.91)
- HORP: Human-Object Relation Priors Guided HOI Detection (CVPR 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.
