| dc.description.abstract | Virtual Reality (VR) is a computer technology that uses computer-generated simulations of a three-dimensional environment that can be interacted with special electronic equipment such as a head-mounted display (HMD). VR users have grown from 5 million in 2015 to over 170 million in 2018. However, authentication systems for virtual reality are relatively underexplored. We propose an authentication method for VR that augments the traditional password-based system with biometric security.
Biometrics is an active area of research in the HCI, pattern recognition, and machine learning communities. Various physiological features such as fingerprint, DNA, iris pattern, and facial recognition are widely used in biometrics. Recently, there has been an interest in using eye movement patterns, gait, and keystroke signature as behavioral bio-metric modalities.
As most VR systems use a head-mounted display, it is difficult to use standard bio-metric authentication systems such as fingerprint, face recognition, etc. Eyes are a natural mode of interaction in the VR domain. Furthermore, inaccuracies due to head movement are eliminated in the VR domain due to head-mounted VR headsets. In this work, we use the human eye as a source of biometric information for VR and augment a password-based authentication system with it. Latent features encoded by behavioral characteristics of individuals are present in eye movement data. This information combined with other authentication methods can create effective security systems for Virtual Reality. We use a head-mounted display with a built-in eye tracker to collect eye movement information. We analyzed the data by extracting eye movement and sketch recognition based features to build a classification system. We obtained results with accuracy values of up to 50%. | |
