Chan, S., Zhang, H., & Nanayakkara, S. (2023, April). Eye Movement Analysis of Human Visual Recognition Processes with Camera Eye Tracker: Higher Mean and Variance of Fixation Duration for Familiar Images. In Extended Abstracts of the 2023 CHI Conference on Human Factors in Computing Systems (pp. 1-8).
Eye movements could provide information on a person’s cognitive context. Previous works showed that human visual recognition processes could be detected via electrooculography (EOG) data and revealed top-ranked eye movement features for detection. However, they lack analysis on how eye movements differ when viewing recognised (familiar) and unrecognised (unfamiliar) images. EOG methods are less prevalent and have limited scalability for detecting other cognitive processes. Our study with 34 participants showed that saccade count and fixation count decreased while the mean and variance of fixation duration increased with image familiarity. We discuss the differences in eye movements when seeing images of different categories and lessons learnt for our next steps to enable implicit interactions using wearable camera-based eye trackers.