Static audio keystroke dynamics

Abstract

In this paper we investigate the accuracy of an identification scheme based on the sound of typing a password. The novelty of this paper lies in the comparison of performance between timing based and audio based keystroke dynamics data in both an authentication and an identification setting.We collected data of 50 people typing the same given password 100 times, divided into 4 sessions of 25 typings, and tested how well the system could recognize the correct typist. When training with data of 3 sessions and testing with the remaining session we achieved a maximal accuracy of 97.3% using cross validation. Repeating this with training with 1 session and testing with the 3 remaining sessions we achieved an accuracy of still 90.6%. The results show the potential of using Audio Keystroke Dynamics information as a way to identify users during log on.