A Keystroke Dynamics Based System for User Identification

Analysis of pattern recognition techniques for in-air signature biometrics

“…Biometric techniques are usually divided into two groups depending on the characteristic used to identify a person, namely physical and behavioural [4]. Physical biometric techniques are based on a physical characteristic preserved in time that a user owns (iris [5], fingerprint [6], hand geometry [7], face [8]) whereas behavioural techniques are related to something that the user is able to repeat in an unique manner (handwriting signature [9], keystroke dynamics [10], gait [11]). In this article, we propose a biometric technique in which a person is authenticated by making his/her handwritten signature in the air (in-air signature) while holding a mobile phone.…”

Biometric personal authentication using keystroke dynamics: A review

“…Hu et al [33] use trigraphs (three consecutively typed keys) also known as Degree of Disorder as features and normalize the distance to find the feature subset. Mariusz et al [34] propose an approach to select the most interesting features and combine them to obtain viable indicator of user's identity. Christopher et al [35] used three stage software design process along with data capture device hardware together with pattern recognition techniques like Bayesian and Discrimination function for classification.…”

On the Comparison of the Keystroke Dynamics Databases

“…Our teams' first documented research on keystroke dynamics is dated back on 2008 [9], where authors presented promising results with the use of simple classification techniques and analyzing only "dwell" and "flight" times. Every single user had to enter three different and independent samples (without any repetitive words), where two of them were 110 keystrokes each (used as reference) and last one had length of about 55 keystrokes (used for validation).…”