A System-Wide Keystroke Biometric System
John DeLuca, Dwight R. Worley, Horace Henry, Pierre Folkes, Ned Bakelman · 2011
The Pace University Keystroke Biometric System (PKBS) uses feature extraction and pattern classification to capture, process, and analyze keystroke timing information to identify typing patterns from limited textbased input. However, keystroke data can come from other sources such as spreadsheets, document-creation programs and web-based applications. This study explores modifying the PBKS to function as a systemwide keylogger and data-analysis suite. An open source keylogger is used to obtain samples of keystroke data which are then analyzed according to the application used. A new converter is designed and implemented to parse raw data files and produce an XML file compatible with the updated PKBS file format. The results demonstrate the performance on each application and determine the system‟s usefulness for identification and authentication.