Visible to the public Keystroke dynamics performance enhancement with soft biometrics

TitleKeystroke dynamics performance enhancement with soft biometrics
Publication TypeConference Paper
Year of Publication2015
AuthorsIdrus, S. Z. Syed, Cherrier, E., Rosenberger, C., Mondal, S., Bours, P.
Conference NameIEEE International Conference on Identity, Security and Behavior Analysis (ISBA 2015)
Date Publishedmar
Keywordsauthentication, behavioural biometric modality, behavioural sciences computing, biometric characteristics, biometrics (access control), classification approach, combination approach, Databases, feature extraction, keystroke dynamics performance enhancement, pubcrawl170115, soft biometrics, Support vector machines, Timing

It is accepted that the way a person types on a keyboard contains timing patterns, which can be used to classify him/her, is known as keystroke dynamics. Keystroke dynamics is a behavioural biometric modality, whose performances, however, are worse than morphological modalities such as fingerprint, iris recognition or face recognition. To cope with this, we propose to combine keystroke dynamics with soft biometrics. Soft biometrics refers to biometric characteristics that are not sufficient to authenticate a user (e.g. height, gender, skin/eye/hair colour). Concerning keystroke dynamics, three soft categories are considered: gender, age and handedness. We present different methods to combine the results of a classical keystroke dynamics system with such soft criteria. By applying simple sum and multiply rules, our experiments suggest that the combination approach performs better than the classification approach with best result of 5.41% of equal error rate. The efficiency of our approaches is illustrated on a public database.

Citation Keyidrus_keystroke_2015