TKCA: a timely keystroke-based continuous user authentication with short keystroke sequence in uncontrolled settings

Abstract Keystroke-based behavioral biometrics have been proven effective for continuous user authentication. Current state-of-the-art algorithms achieved outstanding results in long text or short collected by doing some tasks. It remains a considerable challenge to authenticate users continuously and accurately with keystroke inputs uncontrolled settings. In this work, we propose Timely method Continuous Authentication, named TKCA. integrates the key name two kinds of timing features through an embedding mechanism. And it captures relationship between context keystrokes Bidirectional Long Short-Term Memory (Bi-LSTM) network. We conduct series experiments validate on public dataset - Clarkson II completely natural setting. Experiment show that proposed TKCA achieves performance 8.28% EER when using only 30 2.78% 190 keystrokes.