Quantifying objective and perceived image quality through EEG and eye-tracking

Abstract

Image quality evaluation of electronic displays is inherently subjective. To date, user studies in focus groups and viewers’ self-reports have been the primary sources of feedback used as judgment criteria for the display quality. However, little has been known whether the measured responses to visual stimuli, other than self-reports, reflect the objective and subjective levels of image quality on displays. Here, we used electroencephalogram (EEG) and eye-tracking to investigate whether individuals’ neural and physiological responses to visual stimuli track three objectively discerned levels of image quality, as well as their self-ratings on vividness. Our findings reveal that event-related potentials (ERPs) at 200-300 msec after stimulus onset in the frontocentral brain region, as well as saccade and blink frequencies, significantly tracked (clusterbased permutation p < 0.05) the objective perspective of image quality. In contrast, ERPs at 600 msec in the frontolateral region and saccade peak velocity tracked (cluster-based permutation p < 0.05) the subjective evaluation of image quality. Individuals’ pupil diameter successfully tracked (cluster-based permutation p < 0.05) both objective and subjective perspectives of image quality. These patterns highlight both the shared characteristics and measurable distinctions between objective image quality and subjective vividness ratings. This study demonstrates the effectiveness of EEG and eye-tracking as quantitative tools for assessing objective image quality disparities and the subjective affective nuances of users' visual display experiences, potentially reducing reliance on self-reports.