Research map / Graph learning / Knowledge and temporal graphs
Learning Human-Object Interactions by Graph Parsing Neural Networks
Siyuan Qi, Wenguan Wang, Baoxiong Jia, Jianbing Shen, Song-Chun Zhu
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 graph structure (message passing) on top of node/edge features (ablation) | GPNN · HICO-DET (Full) | mAP (%): 7.88 -> 13.11 | Improved |
| add graph structure on V-COCO (ablation) | GPNN · V-COCO (average) | mAP (%): 28.1 -> 44.0 | Improved |
| add graph structure on CAD-120 object affordance detection (ablation) | GPNN · CAD-120 Detection object affordance | F1-score (%): 20.8 -> 88.8 | Improved |
| learn soft adjacency matrix instead of a constant all-ones graph (ablation) | GPNN · HICO-DET (Full) | mAP (%): 8.75 -> 13.11 | Improved |
| learn soft adjacency matrix instead of constant graph on V-COCO (ablation) | GPNN · V-COCO (average) | mAP (%): 34.3 -> 44.0 | Improved |
| add L1 supervision loss on adjacency matrices (ablation) | GPNN · HICO-DET (Full) | mAP (%): 8.15 -> 13.11 | Improved |
| add L1 graph loss on V-COCO (ablation) | GPNN · V-COCO (average) | mAP (%): 38.4 -> 44.0 | Improved |
| learn graph parsing jointly with message passing (ablation) | GPNN · HICO-DET (Full) | mAP (%): 10.17 -> 13.11 | Improved |
| learn graph parsing jointly with message passing on CAD-120 (ablation) | GPNN · CAD-120 Detection sub-activity | F1-score (%): 79.3 -> 88.9 | Improved |
| use 3 message passing iterations instead of 1 (ablation) | GPNN · HICO-DET (Full) | mAP (%): 11.38 -> 13.11 | Improved |
| use 3 iterations instead of 4 iterations (ablation) | GPNN · V-COCO (average) | mAP (%): 44.0 -> 42.9 | No clear effect |
| use 3 iterations instead of 2 iterations on HICO-DET (ablation) | GPNN · HICO-DET (Full) | mAP (%): 12.37 -> 13.11 | Improved |
Closest papers by meaning
- Cascaded Parsing of Human-Object Interaction Recognition (TPAMI 2021 · similarity 0.94)
- Relation Parsing Neural Network for Human-Object Interaction Detection (ICCV 2019 · similarity 0.90)
- Spatially Conditioned Graphs for Detecting Human-Object Interactions (ICCV 2021 · similarity 0.90)
- Exploring Pose-Aware Human-Object Interaction via Hybrid Learning (CVPR 2024 · similarity 0.87)
- Human-Object Interaction Detection Collaborated with Large Relation-driven Diffusion Models (NeurIPS 2024 · similarity 0.87)
- Learning Human-Object Interaction as Groups (NeurIPS 2025 · similarity 0.86)
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.
