A Multimodal Biometric User Identification System Based on Keystroke Dynamics and Mouse Movements

Abstract

In this work it is shown how the behavioral biometrics allows to strengthen security of a personal computer during casual use. The user does not have to be even aware of verification system running in the background. Unfortunately, short passwords do not supply enough data for keystroke dynamics algorithms to be precise enough to keep the way and level the biometrics system requires. Behavioral biometrics cannot grant such authentication level as the other physiological biometric methods, e.g. fingerprint or retina scan. However, their transparency in analyzing data allows to merge methods into multimodal systems with a minimal cost. The benefit of keystroke dynamics is that it can be easily connected with some other biometric methods, especially with other human input interface devices. In this paper an approach to analyze keystroke dynamics along with mouse movement is presented. Even though both of the features are of behavioral character and hence with low repeatability, the results are good and promising for further research and modification.