CHI Conference on Human Factors in Computing Systems.,
May 2024 · doi: https://doi.org/10.1145/3613905.3651094
Eye gaze has emerged as a promising avenue for implicit authentication/ identification on smartphones, offering the potential for seamless user identification and two-factor authentication. However, a crucial gap exists in understanding eye gaze behaviour specifically during smartphone unlocks. This lack of understanding is magnified by scenarios where users’ faces are not fully visible in front cameras, leading to inaccurate gaze estimation. In this work, we conducted a 24-hour in-the-wild study tracking 21 users’ eye gaze during smartphone unlocks. Our findings highlight substantial eye gaze behaviour variations influenced by authentication methods, physical activity, and environment. Our findings provide insights to enhance and adapt implicit user identification/authentication systems based on gaze tracking on smartphones taking into consideration different users’ behaviour, and environmental effects.
subject terms: Gaze Behaviour, Authentication, Identification, Smartphones, Eye Tracking