Using Motion Forecasting for Behavior-Based Virtual Reality (VR) Authentication. (arXiv:2401.16649v1 [cs.LG])

Task-based behavioral biometric authentication of users interacting in
virtual reality (VR) environments enables seamless continuous authentication by
using only the motion trajectories of the person's body as a unique signature.
Deep learning-based approaches for behavioral biometrics show high accuracy
when using complete or near complete portions of the user trajectory, but show
lower performance when using smaller segments from the start of the task. Thus,
any systems designed with existing techniques are vulnerable while waiting for
future segments of motion …

accuracy arxiv authentication behavioral biometrics biometric biometrics body continuous continuous authentication deep learning environments forecasting high motion near reality signature task virtual virtual reality