|
Activity Number:
|
516
|
|
Type:
|
Invited
|
|
Date/Time:
|
Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statisticians in Defense and National Security
|
| Abstract - #308104 |
|
Title:
|
Statistical Issues in Biometric Authentication
|
|
Author(s):
|
Stephen Fienberg*+ and Sinjini Mitra
|
|
Companies:
|
Carnegie Mellon University and The University of Southern California
|
|
Address:
|
Department of Statistics, Pittsburgh, PA, 15213,
|
|
Keywords:
|
biometrics ; statistical analysis ; ROC curve ; empirical methods ; error rates
|
|
Abstract:
|
A wide variety of biometric authentication systems exist today, based on face, fingerprints and even iris recognition. Following the attacks of 9/11, "biometrics" migrated in use in airports and other public places, and they are about to become an integral part of the US immigration and passport systems. Many biometric systems report surprisingly high error rates and there is an acute lack of rigorous statistical methods for analyzing these. We discuss examples of biometric authentication tools and systems and ways to assess their reliability and accuracy.
|
