Biometric Identification Using User Interactions with Virtual Worlds

Biometric identification systems are being used widely in many applications and security systems all over the world. In this paper, a review of the research in behavioural biometrics is presented and a biometric identification system is proposed which utilises user interactions with virtual worlds for identification. Three game based virtual environments are implemented and three biometric similarity measures namely: distance, Euclidean, and interclass are used to evaluate the system. The proposed system was able to identify users with up to 89 percent of accuracy. The average equal error rates between 29 to 35 percent are achieved. The proposed system can have applications such as identifying users in virtual environments, differentiating humans from machines in online games, and finding users operating multiple accounts in an online system.