Authors: Romain Giot, Christophe Rosenberger
Addresses: GREYC Research Lab, ENSICAEN – Universite de Caen Basse Normandie – CNRS, 14000 Caen, France. ' GREYC Research Lab, ENSICAEN – Universite de Caen Basse Normandie – CNRS, 14000 Caen, France
Abstract: Keystroke dynamics allows to authenticate individuals through their way of typing on a computer keyboard. In this study, we are interested in static shared secret keystroke dynamics (all the users type the same password). We present new soft biometrics information which can be extracted from keystroke typing patterns: the gender of the user. This is the first study, to our knowledge, experimenting such kind of information in the field of keystroke dynamics. We present a method for gender recognition through keystroke dynamics with more than 91% of accuracy, on the tested dataset, and we show the improvement on keystroke dynamics authentication method using such kind of information through pattern and score fusion. We obtain a gain of 20% when using gender information against a classical keystroke dynamics method.
Keywords: gender recognition; soft biometrics; keystroke dynamics; authentication; typing patterns; computer keyboards.
International Journal of Information Technology and Management, 2012 Vol.11 No.1/2, pp.35 - 49
Published online: 01 Dec 2011 *Full-text access for editors Access for subscribers Purchase this article Comment on this article