Research map / Graph learning / Knowledge and temporal graphs
Object-Relation Reasoning Graph for Action Recognition
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 relation-level reasoning to the object-level graph (ablation) | OR2G · Charades, Oracle | mAP: 59.77 -> 62.81 | Helped |
| add the graph aggregating module (ablation) | OR2G · Charades, Oracle | mAP: 62.81 -> 63.28 | Helped |
| add the graph aggregating module (ablation) | OR2G · Charades, Prediction | mAP: 33.60 -> 34.24 | Helped |
| combine spatial and temporal object-graph edges rather than use either alone (ablation) | OG · Charades, Oracle | mAP: 57.95 -> 59.77; 58.94 -> 59.77 | Helped |
| use straight rather than weighted combination of spatial and temporal edges (ablation) | OG · Charades, Oracle | mAP: 59.53 -> 59.77 | Helped |
| use straight rather than weighted combination of spatial and temporal edges (ablation) | OG · Charades, Prediction | mAP: 32.63 -> 32.65 | No clear effect |
| use visual and semantic object features together (ablation) | OG · Charades, Oracle | mAP: 34.56 -> 59.77; 58.78 -> 59.77 | Helped |
| use visual and semantic object features together (ablation) | OG · Charades, Prediction | mAP: 25.51 -> 32.65; 32.36 -> 32.65 | Helped |
| use both semantic and spatial relation features (ablation) | OR2G* · Charades, Oracle | mAP: 60.28 -> 62.81; 62.65 -> 62.81 | Helped |
| use both semantic and spatial relation features (ablation) | OR2G* · Charades, Prediction | mAP: 32.89 -> 33.60; 33.17 -> 33.60 | Helped |
| add the I3D global video feature (ablation) | OR2G · Charades, Prediction | mAP: 34.24 -> 44.91 | Helped |
| add the I3D global video feature (ablation) | OR2G · Charades, Oracle | mAP: 63.28 -> 67.51 | Helped |
Closest papers by meaning
- Graph-Based High-Order Relation Modeling for Long-Term Action Recognition (CVPR 2021 · similarity 0.89)
- Temporal Relational Modeling with Self-Supervision for Action Segmentation (AAAI 2021 · similarity 0.88)
- Improving Action Segmentation via Graph-Based Temporal Reasoning (CVPR 2020 · similarity 0.87)
- Open-Vocabulary Video Relation Extraction (AAAI 2024 · similarity 0.86)
- EGTR: Extracting Graph from Transformer for Scene Graph Generation (CVPR 2024 · similarity 0.85)
- HyperGLM: HyperGraph for Video Scene Graph Generation and Anticipation (CVPR 2025 · similarity 0.85)
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.