Research map / Video / Video segmentation and tracking
Egocentric Activity Prediction via Event Modulated Attention
Yang Shen, Bingbing Ni, Zefan Li, Ning Zhuang
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 asynchronous event-sequence LSTM branch instead of time-series only (ablation) | Time series LSTM · GTEA Gaze | prediction accuracy: 0.581 -> 0.612 | Improved |
| add asynchronous event-sequence LSTM branch instead of time-series only (ablation) | Time series LSTM · GTEA Gaze+ | prediction accuracy: 0.614 -> 0.659 | Improved |
| fuse synchronous and asynchronous features (ablation) | Event sequence LSTM · GTEA Gaze | prediction accuracy: 0.612 -> 0.632 | Improved |
| fuse synchronous and asynchronous features (ablation) | Event sequence LSTM · GTEA Gaze+ | prediction accuracy: 0.659 -> 0.674 | Improved |
| add event modulated attention (ablation) | Fusion LSTMs · GTEA Gaze | prediction accuracy: 0.632 -> 0.648 | Improved |
| add event modulated attention (ablation) | Fusion LSTMs · GTEA Gaze+ | prediction accuracy: 0.674 -> 0.687 | Improved |
| use Rubicon Boundaries re-annotated labels | Time series LSTM recognition · GTEA Gaze+ | recognition accuracy: 0.671 -> 0.686 | Improved |
| use gaze-based LSTM instead of joint training CNNs | Attention based LSTMs motion prediction · GTEA Gaze+ | motion prediction accuracy: 0.576 -> 0.842 | Improved |
| add event modulated attention under Gaussian noise (ablation) | Fusion LSTMs · GTEA Gaze | accuracy drop under noise: drops 14.5% -> 13.1% | Improved |
| use our gaze-event two-stream LSTM instead of Ego ConvNet 2D+3D | prediction · GTEA Gaze | prediction accuracy: 0.514 -> 0.632 | Improved |
Closest papers by meaning
- Gaze-VLM: Bridging Gaze and VLMs through Attention Regularization for Egocentric Understanding (NeurIPS 2025 · similarity 0.88)
- Predicting Gaze in Egocentric Video by Learning Task-dependent Attention Transition (ECCV 2018 · similarity 0.87)
- Self-Regulated Learning for Egocentric Video Activity Anticipation (TPAMI 2021 · similarity 0.86)
- GazeFlow: From Human Gaze Behavior to Generative Egocentric Gaze Prediction (NeurIPS 2026 · similarity 0.84)
- Gaze Beyond the Frame: Forecasting Egocentric 3D Visual Span (NeurIPS 2025 · similarity 0.83)
- Action Scene Graphs for Long-Form Understanding of Egocentric Videos (CVPR 2024 · similarity 0.83)
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.
