Reproduce “GazeIntent: Adapting dwell-time selection in VR interaction with real-time intent modeling” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Anish S. Narkar, Jan J. Michalak, Candace E. Peacock, Brendan David-John
GazeIntent is an adaptive gaze-based selection architecture for Virtual Reality (VR) that dynamically adjusts dwell-time thresholds using real-time user intent prediction. The system enhances interaction efficiency by increasing selection speed and reducing perceived false positives, particularly for returning users and in high-frequency tasks, while maintaining high predictive accuracy with a lightweight model.
Key insights
- Continuous gaze features (e.g., velocity, acceleration, displacement) are significantly more influential in predicting user interaction intent than event-based features, enabling lightweight and high-performance models.
- The intent prediction model's performance remains robust and largely unaffected by changes in user cognitive load, such as varying arithmetic difficulty or the presence of a memory task.
- Personalized intent models and the system's ability to adapt to practice effects (changes in user behavior over time) significantly enhance real-time interaction performance and user preference.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.