Using Motion Forecasting for Behavior-Based Virtual Reality (VR) Authentication

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 authentication behavior biometric biometric authentication biometrics continuous cs.cr cs.lg deep learning environments forecasting near person reality show trajectory virtual virtual reality