Your Eyes Tell: Leveraging Smooth Pursuit to Assess Cognitive Workload
summary
The paper examines using smooth pursuit eye movements to infer users' cognitive workload during interaction. The authors ran a user study (N=20) comparing three pursuit trajectories and two speeds under varying workload, found larger gaze deviations on certain trajectories at higher workload, and trained an SVM that classifies low vs. high workload with 99.5% accuracy and three-level workload with 88.1% accuracy. The work discusses real-time, cognition-aware system use cases that leverage smooth pursuit instead of questionnaires or pupilometry.
AI-written summary · about