Research map / Video / Video-language understanding
Detecting Engagement in Egocentric Video
Yu-Chuan Su, Kristen Grauman
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 interval proposals to frame-wise prediction (ablation) | Ours-frame · UT EE (cross-recorder) | interval F1: 0.533 -> 0.572 | Improved |
| add interval proposals to frame-wise prediction (ablation) | Ours-frame · UT EE (cross-scenario) | interval F1: 0.531 -> 0.553 | Improved |
| add interval proposals to frame-wise prediction (ablation) | Ours-frame · UT EE (cross recorder AND scenario) | interval F1: 0.532 -> 0.544 | Improved |
| add interval proposals to frame-wise prediction (ablation) | Ours-frame · UT EE (cross-recorder) | frame F1: 0.686 -> 0.674 | No clear effect |
| add interval proposals to frame-wise prediction (ablation) | Ours-frame · UT EE (cross-scenario) | frame F1: 0.683 -> 0.665 | No clear effect |
| add interval proposals to frame-wise prediction (ablation) | Ours-frame · UT EE (cross recorder AND scenario) | frame F1: 0.680 -> 0.661 | No clear effect |
| use perfect (ground truth) interval hypotheses instead of proposals | Ours-interval · UT EE (cross-recorder) | interval F1: 0.572 -> 0.868 | Improved |
| replace CNN appearance features with long-term egomotion features | CNN Appearance · UT EE (cross-recorder) | interval F1: 0.486 -> 0.572 | Improved |
| replace CNN appearance features with long-term egomotion features | CNN Appearance · UT EE (cross-scenario) | frame F1: 0.656 -> 0.683 | Improved |
| replace Motion Magnitude with learned long-term motion model | Motion Magnitude · UT EE (cross-recorder) | frame F1: 0.553 -> 0.686 | Improved |
| hold out scenario and recorder (vs recorder only) | Ours-frame · UT EE | frame F1: 0.686 -> 0.680 | No clear effect |
Closest papers by meaning
- EgoEverything: A Benchmark for Human Behavior–Inspired Long-Context Egocentric Video Understanding in AR Environment (ECCV 2026 · similarity 0.88)
- EgoM2P: Egocentric Multimodal Multitask Pretraining (ICCV 2025 · similarity 0.87)
- Where and What: Long-Term Object Tracking in Egocentric Videos (ECCV 2026 · similarity 0.87)
- Action Scene Graphs for Long-Form Understanding of Egocentric Videos (CVPR 2024 · similarity 0.86)
- The Audio-Visual Conversational Graph: From an Egocentric-Exocentric Perspective (CVPR 2024 · similarity 0.86)
- EgoThinker: Unveiling Egocentric Reasoning with Spatio-Temporal CoT (NeurIPS 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.
