Research map / Detection and segmentation / Open-set and few-shot detection
Contextual Heterogeneous Graph Network for Human-Object Interaction Detection
Hai Wang, Wei-shi Zheng, Ling Yingbiao
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 intra-class message to baseline (ablation) | baseline · V-COCO | AP_role: 47.4 -> 52.0 | Improved |
| add inter-class message to baseline (ablation) | baseline · V-COCO | AP_role: 47.4 -> 51.4 | Improved |
| add both intra- and inter-class messages (ablation) | baseline · HICO-DET Default (Full) | mAP: 13.73 -> 17.57 | Improved |
| add both messages (ablation) | baseline · HICO-DET Known Object (Full) | mAP: 16.92 -> 21.00 | Improved |
| add contextual graph attention (ablation) | Proposed heterogeneous graph · V-COCO | AP_role: 49.5 -> 52.7 | Improved |
| add contextual graph attention (ablation) | Proposed heterogeneous graph · HICO-DET Default (Full) | mAP: 15.34 -> 17.57 | Improved |
| add intra-class context in attention (ablation) | Proposed heterogeneous graph · V-COCO | AP_role: 51.2 -> 52.7 | Improved |
| add interactiveness weight (ablation) | Proposed heterogeneous graph · HICO-DET Known Object (Full) | mAP: 19.95 -> 21.00 | Improved |
| use heterogeneous instead of homogeneous graph (intra-class function) (ablation) | Proposed heterogeneous graph · V-COCO | AP_role: 49.3 -> 52.7 | Improved |
| use heterogeneous instead of homogeneous graph (inter-class function) (ablation) | Proposed heterogeneous graph · HICO-DET Default (Full) | mAP: 12.6 -> 17.5 | Improved |
| add intra-class message (ablation) | Proposed · V-COCO complex scenes | AP_role: 54.7 -> 55.1 | Improved |
| add intra-class message (ablation) | Proposed · HICO-DET simple scenes | mAP: 15.7 -> 15.7 | No clear effect |
Closest papers by meaning
- Exploiting Scene Graphs for Human-Object Interaction Detection (ICCV 2021 · similarity 0.91)
- Learning Human-Object Interactions by Graph Parsing Neural Networks (ECCV 2018 · similarity 0.90)
- Detecting Human-Object Interactions via Functional Generalization (AAAI 2020 · similarity 0.89)
- Learning Human-Object Interaction as Groups (NeurIPS 2025 · similarity 0.89)
- Learning to Detect Human-Object Interactions With Knowledge (CVPR 2019 · similarity 0.89)
- ContextHOI: Spatial Context Learning for Human-Object Interaction Detection (AAAI 2025 · similarity 0.87)
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.
