Leap motion controller for authentication via hand geometry and gestures

Abstract

The Leap Motion controller is a consumer gesture sensor aimed to augment a user’s interactive experience with their computer. Using infrared sensors, it is able to collect data about the position and motions of a user’s hands. This data allows the Leap to be used as an authentication device. This study explores the possibility of performing both login as well as continuous authentication using the Leap Motion device. The work includes classification of static data gathered by the Leap Motion using trained classifiers, with over 99% accuracy. In addition, data was recorded from the users while utilizing the Leap Motion to read and navigate through Wikipedia pages. A template was created using the user attributes that were found to have the highest merit. The algorithm found when matching the template to the users newly collected data, the authentication provided an accuracy of over 98%, and an equal error rate of 0.8% even for a small number of attributes. This study demonstrates that the Leap Motion can indeed by used successfully to both authenticate users at login as well as while performing continuous activities. As the Leap Motion is an inexpensive device, this raises the potential of using its data in the future for authentication instead of traditional keyboard passwords.