Decoding Visual Fatigue Through Pupil Dynamics

This study delves into the intricate relationship between visual fatigue and pupil parameters, introducing pupil metrics as an innovative approach for visual fatigue assessment. We explored the associations between eight key pupil metrics (including average pupil diameter, pupil fluctuation, pupil adaptation speed, percentage of eye closure, blink duration, blink frequency, Saccade duration, and peak Saccade speed) and subjective ratings of visual fatigue to pinpoint the most indicative metrics for characterizing visual fatigue. Under challenging lighting conditions, we engaged seven volunteers in a visual ergonomics experiment. The outcomes underscored that six of the pupil metrics strongly aligned with the subjective trends of visual fatigue scores. Leveraging the SVM machine learning algorithm and focusing on these six metrics, we achieved a predictive accuracy of 92.86% for visual fatigue, marking a 7.15% improvement over predictions made using all metrics. Overall, our findings substantiate the robustness of these six pupil parameters in evaluating visual fatigue. Furthermore, we crafted a predictive model rooted in these metrics, presenting a potent tool for the detection of visual fatigue.