Research map / Detection and segmentation / Open-set and few-shot detection
Amplifying Key Cues for Human-Object-Interaction Detection
Yang Liu, Qingchao Chen, Andrew Zisserman
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 fine-grained spatial configuration (ablation) | Baseline · HICO-DET Default (Full) | mAP: 14.8 -> 17.6 | Improved |
| add context encoder (ablation) | Baseline · HICO-DET Default (Full) | mAP: 14.8 -> 16.2 | Improved |
| add motion encoder (plausible flow) (ablation) | Baseline · HICO-DET Default (Full) | mAP: 14.8 -> 16.5 | Improved |
| combine all modules (full FCMNet) (ablation) | Baseline · HICO-DET Default (Full) | mAP: 14.8 -> 20.4 | Improved |
| use human parts instead of human skeleton (ablation) | Base · HICO-DET Default (Full) | mAP: 15.8 -> 17.1 | Improved |
| add human parts and object mask together (ablation) | Baseline · HICO-DET Default (Full) | mAP: 14.8 -> 17.6 | Improved |
| encode context via word2vec instead of visual appearance (ablation) | Base+Both · HICO-DET Default (Full) | mAP: 15.2 -> 16.2 | Improved |
| add global context only (word2vec) (ablation) | Baseline · HICO-DET Default (Full) | mAP: 14.8 -> 15.1 | Improved |
| gate-and-memory fusion instead of late fusion (ablation) | FCMNet · HICO-DET Default (Full) | mAP: 18.1 -> 20.4 | Improved |
| gate-and-memory fusion instead of concatenation (ablation) | FCMNet · HICO-DET Default (Full) | mAP: 17.9 -> 20.4 | Improved |
| gate-and-memory fusion instead of attention fusion (ablation) | FCMNet · HICO-DET Default (Full) | mAP: 19.7 -> 20.4 | Improved |
Closest papers by meaning
- Learning Human-Object Interaction Detection Using Interaction Points (CVPR 2020 · similarity 0.93)
- Detecting Human-Object Interactions via Functional Generalization (AAAI 2020 · similarity 0.92)
- LINK: Learning Instance-level Knowledge from Vision-Language Models for Human-Object Interaction Detection (ICLR 2026 · similarity 0.91)
- Learning Human-Object Interaction as Groups (NeurIPS 2025 · similarity 0.90)
- Exploring Pose-Aware Human-Object Interaction via Hybrid Learning (CVPR 2024 · similarity 0.90)
- 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.
