Research map / Detection and segmentation / Open-set and few-shot detection
Discovering Human Interactions With Large-Vocabulary Objects via Query and Multi-Scale Detection
Suchen Wang, Kim-Hui Yap, Henghui Ding, Jiyan Wu, Junsong Yuan, Yap-Peng Tan
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 parallel multi-scale feature maps (ablation) | DETR-like baseline · SWiG-HOI | APs: 1.24 -> 2.11 | Improved |
| add parallel multi-scale feature maps (ablation) | DETR-like baseline · SWiG-HOI | AP: 5.16 -> 5.28 | Improved |
| add cascaded detection (ablation) | multi-scale model · SWiG-HOI | AP: 5.28 -> 5.70 | Improved |
| couple object classification with action prediction (ablation) | multi-scale cascade model · SWiG-HOI | AP: 5.70 -> 7.31 | Improved |
| couple action prediction with object classification (action ablated) (ablation) | Ours · SWiG-HOI | AP: 1.61 AP drop | Improved |
| use HOI-specific joint detection instead of generic Faster RCNN with RFS | Faster RCNN · SWiG-HOI | AR (top 10 boxes): 38.16 -> 57.96 | Improved |
| use joint query-based detection instead of PPDM/DIRV | one-stage HOI detectors · SWiG-HOI | mAP: 3.17 -> 7.98 | Improved |
| use joint detection instead of separate detector and matching | PPDM · SWiG-HOI | inference speed (ms): 237 -> 93 | Improved |
| apply query method to novel interactions | JSR · SWiG-HOI | mAP-novel: 2.34 -> 2.64 | Improved |
| apply query method to hand-object interaction | PPDM · 100DOH-derived | AP: 25.5 -> 28.0 | Improved |
| large-vocabulary techniques (EQL) vs RFS | Faster RCNN · SWiG-HOI | AP: 6.12 -> 5.74 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DIRV — DIRV: Dense Interaction Region Voting for End-to-End Human-Object Interaction Detection (AAAI 2021) · better on 6 of 6 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 0.92)
- Exploring the Potential of Large Foundation Models for Open-Vocabulary HOI Detection (CVPR 2024 · similarity 0.92)
- QPIC: Query-Based Pairwise Human-Object Interaction Detection With Image-Wide Contextual Information (CVPR 2021 · similarity 0.92)
- LINK: Learning Instance-level Knowledge from Vision-Language Models for Human-Object Interaction Detection (ICLR 2026 · similarity 0.92)
- Learning Human-Object Interaction Detection Using Interaction Points (CVPR 2020 · similarity 0.91)
- Learning Human-Object Interaction as Groups (NeurIPS 2025 · similarity 0.91)
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.
