Biométrie douce pour la dynamique de frappe au clavier
Syed Zulkarnain Syed Idrus · HAL (Le Centre pour la Communication Scientifique Directe) · 2014
At present, there are a variety of usages of biometric techniques for a lot of distinct purposes including physical access control, attendance monitoring, e-payment and others. In order to avoid password-based authentication problems, this research focuses on biometric authentication and we propose to use keystroke dynamics. The reduced performances of keystroke dynamics could be revealed by the higher intra-class variability in the users’ habits. One way to take care of this variability is to take into consideration additional information in the determination process.We propose several contributions in order to enhance the performance recognition of keystroke dynamics systems with our novel methods. For starters, we made our personal dataset, which is a new biometric benchmark database known as ‘GREYC-NISLAB Keystroke’ to satisfy the goal of the thesis, where we had made our own data collection of 110 users, both France and Norway. This new benchmark database is accessible to the international scientific community and features some profiling information about end users: the way of typing (one hand or two hands), gender, age and handedness. In order to increase the system performance, subsequently, we study the biometric fusion with keystroke dynamics. Finally, we present an improvement of user recognition, by combining the authentication process with soft criteria. The outcomes of the experiments display the benefits of the proposed approaches.