Assessing Visual Fatigue of Welding Mask Wearers Through Infrared Imaging-Based Ocular Feature Analysis

Electronic devices have become indispensable to individuals today. Looking at these electronic displays for extended periods causes visual fatigue. Herein, we propose three metrics for objective and quantitative measurement of visual fatigue: blink frequency, blink duration, and pupil accommodation speed. After taking eye images using an infrared camera, binarization and contour detection were performed to localize the pupil. The diameter of the localized pupil region, defined as pupil size, was calculated based on the number of black pixels in the pupil region. The blink frequency, blink duration, and pupil accommodation speed were measured by analyzing successive pupil sizes. To obtain pupil accommodation speed, inflection points were extracted after removing high frequency elements in the frequency domain. After inducing visual stimulation on 10 participants by welding sparks for 30 minutes, two incremental fatigue periods were observed. Blink frequency and duration increased in a similar manner, and pupil accommodation speed decreased in the sections. Therefore, it can be confirmed that as the degree of visual fatigue increases, blink frequency and duration increase, and pupil accommodation speed decreases. Moreover, we confirmed that when eyes become tired, the tendency of blink frequency and duration is opposite to that of the pupil accommodation speed.