Event-related pupillary response-based authentication system using eye-tracker add-on augmented reality glasses for individual identification

• Published in: Frontiers in physiology Vol. 15; p. 1325784
• Main Authors: Park, Sangin, Ha, Jihyeon, Kim, Laehyun
• Format: Journal Article
• Language: English
• Published: Switzerland Frontiers Media S.A 13.08.2024
• Subjects: augmented reality; authentication; biometrics; event-related potential; event-related pupillary response; Physiology; biometrics; event-related pupillary response; augmented reality; event-related potential; authentication
• ISSN: 1664-042X; 1664-042X
• DOI: 10.3389/fphys.2024.1325784

This study aimed at developing a noncontact authentication system using event-related pupillary response (ErPR) epochs in an augmented reality (AR) environment. Thirty participants were shown in a rapid serial visual presentation consisting of familiar and unknown human photographs. ErPR was compared with event-related potential (ERP). ERP and ErPR amplitudes for familiar faces were significantly larger compared with those for stranger faces. The ERP-based authentication system exhibited perfect accuracy using a linear support vector machine classifier. A quadratic discriminant analysis classifier trained using ErPR features achieved high accuracy (97%) and low false acceptance (0.03) and false rejection (0.03) rates. The correlation coefficients between ERP and ErPR amplitudes were 0.452-0.829, and the corresponding Bland-Altman plots showed a fairly good agreement between them. The ErPR-based authentication system allows noncontact authentication of persons without the burden of sensor attachment via low-cost, noninvasive, and easily implemented technology in an AR environment.