Publication: Development of a web-based typing biometrics system for user authentication
Loading...
Date
2007-03-01
Authors
Lee, Soo San
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Password authentication is the most prevalently used identification system in web-based or traditional computer access terminals. Unfortunately, this identification system seems to be inadequate to protect resources because password can be easily cracked, guessed, stolen or deliberately shared. Keystroke dynamic, an automated biometrics method that analyzes the way a person types on a keyboard, is emerging as a key aspect in new security systems. In this project, an online behavioral biometric verification system is developed by employing keystroke dynamic to improve upon the security level provided by password matching while greatly reducing the risk of dictionary-based attacks. This web-based typing biometric authentication system can be customized for multiple Internet-based applications requiring secure authentication. The system uses no specialized equipment, requiring only an Internet capable computer with a keyboard. 80 users’ typing patterns were collected for evaluation purpose. All users were requested to enter the same password. Fuzzy ARTMAP (FAM) neural network is used to analyze and categorize the keystroke dynamics of users. The results were very encouraging, with accuracy rate of 96.88%, False Acceptance Rate (FAR) of 2.5% and False Rejection Rate (FRR) of 3.75%. The results demonstrated that the proposed methods are promising.