Touch Biometrics for Unobtrusive Continuous Authentication onSmartphones
Year: 2015
Type: dissertation
Abstract: Mobile devices such as smartphones have until now been protected by \ntraditional authentication methods, including passwords or PINs. These authentication \nmechanisms are di cult to remember and often disabled, leaving \nthe device vulnerable if stolen. This dissertation presents an unobtrusive, \ncontinuous authentication system based on biometric data collected from a \ntouchscreen. A model is presented for implementation of such a touch biometric \nsystem on a smartphone. This model is evaluated by conducting an \nexperiment simulating real-world touch interaction. Touch data was collected \nfrom 30 participants during normal phone use. This data was analysed using \ntwo classi cation algorithms to determine the accuracy of matching touch \ndata to users. The results show that touch data is distinct between users and \ncould be used in authentication without disrupting normal touch interaction. \nFurthermore, touch data could be used in its raw form without signi cant \npre-processing.
Author: Christina J. Kroeze
Institution: University of Pretoria
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W1030499059
