Comparing Webcam-Based Eye Tracking and Dedicated Eye Tracking for Educational Research

Abstract

Using webcams for eye tracking could enable eye-tracking in classrooms, which allows teachers to tailor their instructions and support to the individual learning process of each student. To advance the eye-tracking research field towards that goal, more information is needed on the quality of webcam-based eye-tracking data and the validity of measuring specific relations between specific gaze patterns and cognitive processes. This study aims to contribute by quantifying the relation between data collected with a dedicated eye tracker and a webcam. Therefore, this study compared webcam-based eye-tracking data and dedicated eye-tracking data in terms of data quality, saccade detection, and performed a correlational analysis on the gaze patterns detected in both types of data. The results show that data quality, mainly in terms of precision and sampling rate, is inferior to that of a dedicated eye tracker. Saccade detection is feasible with the custom velocity-based saccade detection algorithm, even with the lower data quality, especially for longer, horizontal saccades. Vertical, shorter saccade detection is possible but has limits; therefore, the recommended minimum stimulus size is 17.5 x 8.5 cm. A significant correlation was found between the gaze patterns detected in the webcam-based eye-tracking data and the dedicated eye-tracking data.