Research map / Video / Video-language understanding
Detecting Human-Object Relationships in Videos
Jingwei Ji, Rishi Desai, Juan Carlos Niebles
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 InterTx-2 to the object branch (over IntraTx only) (ablation) | HORT · Action Genome SGDet image | recall@20: 36.03 -> 36.92 | Improved |
| add InterTx-3 to the object branch (ablation) | HORT · Action Genome SGDet image | recall@20: 36.92 -> 37.19 | Improved |
| add InterTx-2 to the relation branch (over IntraTx only) (ablation) | HORT · Action Genome SGDet image | recall@20: 35.93 -> 36.36 | Improved |
| add InterTx-3 to the relation branch (ablation) | HORT · Action Genome SGDet image | recall@20: 36.36 -> 37.19 | Improved |
| add InterTx-2 to the object branch (over IntraTx only) (ablation) | HORT · Action Genome SGDet image | recall@50: 46.10 -> 47.51 | Improved |
| add InterTx-3 to the relation branch (ablation) | HORT · Action Genome SGDet image | recall@50: 46.55 -> 47.76 | Improved |
| add the transformer object branch (object scores) (ablation) | HORT · Action Genome SGDet | recall | Improved |
| add 3-D positional encoding versus none (ablation) | HORT · Action Genome SGDet | recall@20: 36.21 -> 37.19 | Improved |
| add 3-D positional encoding versus none (ablation) | HORT · Action Genome SGDet | recall@50: 46.37 -> 47.76 | Improved |
| use spatial xy encoding instead of temporal t only (ablation) | HORT · Action Genome SGDet | recall@20: 36.40 -> 36.82 | Improved |
| add temporal encoding to xy encoding (ablation) | HORT · Action Genome SGDet | recall@50: 47.25 -> 47.76 | Improved |
| add temporal reasoning via 5-frame logit averaging to RelDN | RelDN · Action Genome SGDet image | recall@20: 35.80 -> 35.85 | No clear effect |
Closest papers by meaning
- Pose-Aware Multi-Level Feature Network for Human Object Interaction Detection (ICCV 2019 · similarity 0.89)
- HCVRD: A Benchmark for Large-Scale Human-Centered Visual Relationship Detection (AAAI 2018 · similarity 0.89)
- SportsHHI: A Dataset for Human-Human Interaction Detection in Sports Videos (CVPR 2024 · similarity 0.88)
- Object Level Visual Reasoning in Videos (ECCV 2018 · similarity 0.88)
- Video Relationship Reasoning Using Gated Spatio-Temporal Energy Graph (CVPR 2019 · similarity 0.87)
- Interaction-centric Spatio-Temporal Context Reasoning for Multi-Person Video HOI Recognition (ECCV 2024 · 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.
