Soft Biometrics for Keystroke Dynamics

Abstract

Keystroke dynamics is a viable and practical way as an addition to security for identity verification. It can be combined with passphrases authentication resulting in a more secure verification system. This paper presents a new soft biometric approach for keystroke dynamics. Soft biometrics traits are physical, behavioral or adhered human characteristics, which have been derived from the way human beings normally distinguish their peers (e.g. height, gender, hair color etc.). Those attributes have a low discriminating power, thus not capable of identification performance. Additionally, they are fully available to everyone which makes them privacy-safe. Thus, in this study, it consists of extracting information from the keystroke dynamics templates with the ability to recognise the hand(s) used (i.e. one/two hand(s)); the gender; the age category; and the handedness of a user when he/she types a given password or passphrase on a keyboard. Experiments were conducted on a keystroke dynamics database of 110 users and our experimental results show that the proposed methods are promising.

ACO-Random Forest Approach to Protect the Kids from Internet Threats through Keystroke

“…Giot et al [12] show that it is possible to detect the gender and they reported the accuracy rate more than 90% using typing style. Idrus et al [11] show that it is possible to identify the gender, age group, handedness and one or two hands used while typing and they reported the accuracy rate very close to 90%. Uzun et al [3] show that it is possible to identify the child group and adults through typing pattern and they obtained the accuracy more than 90% for the simple familiar Turkish text.…”

Age Detection Through Keystroke Dynamics from User Authentication Failures

“…Finally, Idrus et al (2013) in a research similar to this paper attempt to extract information from keystroke dynamics templates with the ability to recognise, amongst others, the age category of a user when he or she types a given password or passphrase on a keyboard. They collected the data from 110 users who were asked to type 5 phrases of a length between 17 to 24 characters, using an SVM classifier, separated the users into two age classes, those who were under 30 and those who were over 30, and achieved a classification accuracy between 65% and 82%.…”

A Survey on Keystroke Dynamics Biometrics: Approaches, Advances, and Evaluations

“…With a boost by the gender score computed from the keystroke latency features, they further improved the keystroke dynamics authentication accuracy from an EER of 10.65% to 8.45%, achieving a 20% error reduction. Most recently, keystroke dynamics has been studied to extract not only gender information, but also other soft biometric traits including age category, single or two-handed usage, and left or right-handedness [49,48,50]. Encouraging results using support vector machines have shown promise in keystroke biometrics as a soft biometrics when evaluated on the GREYC-NISLAB Keystroke Dynamics Soft Biometrics Dataset of 110 users with 100 sample passphrases per user (Section 1.4.2.11).…”

“…There were 10 repetitions per phrase per subject for each way of typing. Some performance analysis on the dataset was published in [49,48]. This dataset can be accessed from the following link: http://www.epaymentbio metrics.ensicaen.fr/images/pdf/greyc-nislab%20keystroke%20benchmark%2 0dataset.xls.…”