Research map / Video / Action recognition
Ordered Atomic Activity for Fine-grained Interactive Traffic Scenario Understanding
Nakul Agarwal, Yi-Ting Chen
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 appearance graph to the motion-only model (ablation) | Ordered atomic activity GCN framework · OATS multilabel atomic activity recognition | mAP: 14.55 -> 26.70 | Improved |
| add motion graph to the appearance-only model (ablation) | Ordered atomic activity GCN framework · OATS multilabel atomic activity recognition | mAP: 19.99 -> 26.70 | Improved |
| use object-aware graph models instead of video-level backbones | multilabel atomic activity recognition · OATS multilabel atomic activity recognition | mAP: 14.32 -> 20.26 | Improved |
| late fusion of appearance and motion graph features over other fusion schemes (ablation) | Ordered atomic activity GCN framework · OATS | mAP | Improved |
| proposed model versus MuCon for activity order | activity order prediction · OATS activity order prediction (without start/end tokens) | matching score: 15.20 -> 16.08 | No clear effect |
| proposed framework versus ARG on Collective Activity | Ordered atomic activity GCN framework · Collective Activity (group activity) | group action accuracy: 85.28% -> 86.58% | Improved |
Closest papers by meaning
- Action-slot: Visual Action-centric Representations for Multi-label Atomic Activity Recognition in Traffic Scenes (CVPR 2024 · similarity 0.89)
- Trajeglish: Traffic Modeling as Next-Token Prediction (ICLR 2024 · similarity 0.86)
- WOMD-Reasoning: A Large-Scale Dataset for Interaction Reasoning in Driving (ICML 2025 · similarity 0.84)
- Learning to Evolve Scenes: Reasoning about Human Activities with Scene Graphs (NeurIPS 2026 · similarity 0.84)
- TUMTraf VideoQA: Dataset and Benchmark for Unified Spatio-Temporal Video Understanding in Traffic Scenes (ICML 2025 · similarity 0.84)
- Language-Driven Interactive Traffic Trajectory Generation (NeurIPS 2024 · similarity 0.84)
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.
