Keystroke dynamics: Characteristics and opportunities

Abstract

Significant research into the feasibility of keystroke dynamics as a potential biometric authentication method has taken place since the advent of computers. The studies have progressed from examining typing patterns on desktop keyboards using statistical pattern classifiers to mobile keyboards using neural networks as a pattern classifier. The studies do not have a unifying method of comparing results, which limits comparison between the methods presented. Without the ability to compare studies within research areas, future study is limited in its ability to provide important modifications to the work in question. This paper reviews a representative subset of the current research in keystroke dynamics, and provides recommendations on the potential direction of future work in this area. This will provide a set of guidelines that can be followed by researchers intending to do further work in the area of keystroke dynamics.