Research map / Video / Action recognition
Deep Dual Relation Modeling for Egocentric Interaction Recognition
Haoxin Li, Yijun Cai, Wei-Shi Zheng
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 the interactive LSTM with both symmetrical blocks and the relation branch instead of concatenating features (ablation) | dual relation model · PEV | accuracy: 60.8 -> 64.2 | Improved |
| use the interactive LSTM with both symmetrical blocks and the relation branch instead of concatenating features (ablation) | dual relation model · NUS(first h-h) | accuracy: 77.9 -> 80.2 | Improved |
| add symmetrical LSTM blocks to feature concatenation (ablation) | dual relation model · PEV | accuracy: 60.8 -> 62.7 | Improved |
| add symmetrical LSTM blocks to feature concatenation (ablation) | dual relation model · NUS(first h-h) | accuracy: 77.9 -> 78.1 | Improved |
| add the relation LSTM branch to feature concatenation (ablation) | dual relation model · PEV | accuracy: 60.8 -> 63.0 | Improved |
| add the relation LSTM branch to feature concatenation (ablation) | dual relation model · NUS(first h-h) | accuracy: 77.9 -> 79.0 | Improved |
| add local appearance features instead of using global appearance alone (ablation) | appearance feature model · PEV | accuracy: 40.7 -> 43.2 | Improved |
| add local appearance features instead of using global appearance alone (ablation) | appearance feature model · NUS(first h-h) | accuracy: 63.8 -> 65.1 | Improved |
| add relation modeling to concatenated global and local appearance (ablation) | appearance feature model · PEV | accuracy: 44.2 -> 45.9 | Improved |
| add relation modeling to concatenated global and local appearance (ablation) | appearance feature model · NUS(first h-h) | accuracy: 66.8 -> 68.2 | Improved |
| add relation modeling to concatenated global and local motion (ablation) | motion feature model · PEV | accuracy: 53.2 -> 56.6 | Improved |
| add relation modeling to concatenated global and local motion (ablation) | motion feature model · NUS(first h-h) | accuracy: 73.4 -> 75.0 | Improved |
Closest papers by meaning
- Egocentric Action Recognition by Automatic Relation Modeling (TPAMI 2022 · similarity 0.94)
- Action Scene Graphs for Long-Form Understanding of Egocentric Videos (CVPR 2024 · similarity 0.87)
- The Audio-Visual Conversational Graph: From an Egocentric-Exocentric Perspective (CVPR 2024 · similarity 0.87)
- ANNEXE: Unified Analyzing, Answering, and Pixel Grounding for Egocentric Interaction (CVPR 2025 · similarity 0.86)
- Do Egocentric Video-Language Models Truly Understand Hand-Object Interactions? (ICLR 2025 · similarity 0.86)
- EgoChoir: Capturing 3D Human-Object Interaction Regions from Egocentric Views (NeurIPS 2024 · 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.
