ScholarWorks@성균관대학교

초록

We propose an implicit user authentication system for virtual reality (VR), leveraging natural interactions such as grabbing, pointing, and typing. A key challenge in VR authentication is the limited user data available during initial registration, along with the need to pre-collect attacker data for binary-classification settings. In this paper, we address these limitations and develop a practical implicit user authentication solution. We conducted a study with 16 participants, achieving an average Equal Error Rate (EER) of 6.90%, with typing interactions demonstrating a notably low EER of 3.86%. This approach highlights the potential for seamless and unobtrusive authentication in VR environments. © 2025 Copyright held by the owner/author(s).

키워드