Secure behavioral biometric authentication with leap motion

Abstract

In this paper we examine the effectiveness of user authentication using biometrics and behavioral motion data captured by the Leap Motion sensor. The biometrics data is derived from the user's hand and the behavioral motion data is generated when the user signs his or her signature using his or her hand in front of the sensor. We have developed a prototype system to collect experiment data from 10 participants and used the data to analyze the accuracy and effectiveness of our authentication method. The experimental results are measured by FAR, FRR, and EER. For the hand biometrics data involving 17 genuine hand samples and 162 attacking ones for each of the 10 users, the system has achieved an average EER of 34.80%. For the behavioral signature motion data involving 17 genuine samples and 262 attacking samples for each of the 10 users, the system has achieved an average EER of 3.75% Our study indicates that behavioral biometrics with Leap Motion is a viable authentication approach.