Investigating, Implementing and Evaluating Client-Side Keystroke Analysis User Authentication for Web Sites |
In today's electronic information society security is certainly the challenge of the 21st century. Driven by the need to counteract ever more determined cyber criminals, focus is turning upon biometric security as a means of ensuring protection and privacy of the most sensitive information. Keystroke dynamics and the analysis of the way people type their password is one method drawing significant attention because of its non-invasive nature and lack of requirements for expensive additional hardware. The majority of research has been executed in a local area network environment but this paper examines the possibility of implementing a solution for web sites and whether refinement of comparison data over time would lead to increasing improvement. Although the web site solution is not conclusive further investigation into profile refinement indicates that this may not have a significant impact on authentication rates. The observed typing characteristics of subjects at the beginning, throughout and at the end of the testing period offer little evidence for implementing a profile refinement strategy.
Hocking C, Haskell-Dowland PS (Dowland PS)