Research map / Label-efficient and robust learning / Recognition and long-tail
A Domain Based Approach to Social Relation Recognition
Qianru Sun, Bernt Schiele, Mario Fritz
Check 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 |
|---|---|---|---|
| Use semantic attribute features instead of end-to-end finetuned features | Social relation recognition on PIPA · PIPA AC splits | accuracy: 46.2% -> 57.2% | Improved |
| Use semantic attribute features instead of end-to-end finetuned features | Social domain recognition on PIPA · PIPA AC splits | accuracy: 59.0% -> 67.8% | Improved |
| Use semantic attribute features instead of end-to-end finetuned features | Domain prediction for unseen relations · PIPA SR splits | accuracy: 18.5% -> 33.3% | Improved |
| Add body attribute features to head attribute features (ablation) | All attributes, SVM · PIPA AC splits | accuracy: 44.8% -> 57.2% | Improved |
| Add body attribute features to head attribute features (ablation) | All attributes, SVM · PIPA AC splits | accuracy: 59.4% -> 67.8% | Improved |
| Add head attribute features to body attribute features (ablation) | All attributes, SVM · PIPA AC splits | accuracy: 57.2% -> 57.2% | No clear effect |
| Add head attribute features to body attribute features (ablation) | All attributes, SVM · PIPA AC splits | accuracy: 67.7% -> 67.8% | No clear effect |
| Use body attributes only rather than head attributes only (ablation) | Attribute-based relation recognition · PIPA AC splits | accuracy: 44.8% -> 57.2% | Improved |
| Use body attributes only rather than head attributes only (ablation) | Attribute-based domain recognition · PIPA AC splits | accuracy: 59.4% -> 67.7% | Improved |
| Use activity features alone rather than all attributes (ablation) | Attribute-based relation recognition · PIPA AC splits | accuracy: 52.4% -> 57.2% | Hurt |
| Use activity features alone rather than all attributes (ablation) | Attribute-based domain recognition · PIPA AC splits | accuracy: 64.5% -> 67.8% | Hurt |
| Use clothing features alone rather than all attributes (ablation) | Attribute-based relation recognition · PIPA AC splits | accuracy: 51.4% -> 57.2% | Hurt |
Closest papers by meaning
- Dual-Glance Model for Deciphering Social Relationships (ICCV 2017 · similarity 0.87)
- Visual Social Relationship Recognition (IJCV 2020 · similarity 0.87)
- Social Relation Recognition From Videos via Multi-Scale Spatial-Temporal Reasoning (CVPR 2019 · similarity 0.87)
- SocialGPT: Prompting LLMs for Social Relation Reasoning via Greedy Segment Optimization (NeurIPS 2024 · similarity 0.85)
- Behavioral Geometric Supervision Aligns Video Foundation Models with Human Social Perception (NeurIPS 2026 · similarity 0.85)
- Modeling Multimodal Social Interactions: New Challenges and Baselines with Densely Aligned Representations (CVPR 2024 · similarity 0.83)
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.
