Keystroke analysis for thumb-based keyboards on mobile devices

Abstract

The evolution of mobile networking has opened the door to a wide range of service opportunities for mobile devices, increasing at the same time the sensitivity of the information stored and access through them. Current PIN-based authentication has proved to be an insufficient and an inconvenient approach. Biometrics have proven to be a reliable approach to identity verification and can provide a more robust means of security, as they rely upon personal identifiers. Amongst various biometric techniques available, keystroke analysis combines features that can offer a cost effective, non-intrusive and continuous authentication solution for mobile devices. This research has been undertaken in order to investigate the performance of keystroke analysis on thumb-based keyboards that are being widely deployed upon PDA's and Smartphone devices. The investigation sought to authenticate users whilst typing text messages, using two keystroke characteristics, the inter-keystroke latency and hold-time. The results demonstrate the approach to be promising, achieving an average EER=12.2% with the inter-keystroke latency based upon 50 participants. Uniquely to this tactile environment however, the hold-time characteristic, did not prove to be a reliable feature to be utilised. © 2007 International Federation for Information Processing.