Recover Password  New User
Undergraduate Research Project Management System

Multiple Factor Authentication Using Gaze Tracking and Iris Scanning

Status Current
Seeking Researchers Yes
Start Date 03/01/2006
End Date 05/30/2013
Funding Source Chancellor's Research Fund, National Science Foundation
Funding Amount 140,000
Community Partner
Related Course
Last Updated 09/21/2011 10:40PM
Keywords eye tracking, authentication

People

Faculty
  Bogdan Hoanca, Kenrick Mock

Student Researchers

Abstract

With the advent of ubiquitous mobile networks comes an increase in the use of these networks to access personal information from public places. Of critical importance is the security of these systems to prevent the disclosure of sensitive information. A key component of systems security is user authentication - the process of certifying whether a user is who he or she claims to be. The most common form of authentication is for the computer system to match a text-based username and password entered by the user with the corresponding username and password stored on system. An increasingly popular form of biometric authentication is to match a user's fingerprint with a fingerprint stored on the system. The advantage of the biometric approach is there are no passwords to remember. However, if the biometric is stolen or faked then there is little recourse for the user. For example, a user can't change their fingerprint like they can change their text-based password.

In this project we propose a novel biometric-based authentication scheme that combines iris scanning with gaze tracking. The hybrid scheme is able to uniquely identify a user, is highly resistant to fakery, and is simple and elegant for the end user to authenticate. Iris scanning is similar to fingerprint recognition except it identifies the subject by unique structures that develop in the iris of the eye. One weakness of iris scanning is that it can be foiled using a printed iris image. To offset this weakness, we propose to incorporate gaze tracking. The same camera system used to scan the iris can also be used to detect the location of the user's gaze with high accuracy. This type of system has been used for years by paralyzed and disabled users to control a mouse cursor with their eye. While individual users exhibit different gaze patterns when looking at the same image, these patterns are not unique enough to be used for identification. However, gaze tracking can be used with high accuracy to determine if the eye is live and not a printed image. We propose to utilize the strength of gaze tracking to determine whether the eye is live and combine the result with the strength of iris scanning which is to uniquely identify the authenticating user. The resulting system will be extremely easy to use - the authenticating user merely needs to look at some objects on the computer screen. For even stronger authentication a traditional or graphical password can be incorporated. In the gaze tracking version of the graphical password the user must look at specific locations on the screen in order to authenticate.

Shared Project Files (e.g. papers, presentations)

File name Description Uploaded by