A note on authentication accuracy with multiple biometric images

Kensuke Baba, Serina Egawa

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper conducts an analysis on accuracy of biometric authentication. By increasing the number of biometric images for authentication, authentication accuracy is expected to be improved. However, the relation between the number of images and accuracy is not trivial. This paper considers simple algorithms for verification and identification with multiple biometric images for each person. The algorithms are based on the ideas of a majority vote and the mean of similarities for treating results of comparisons with multiple images. The effects of the number of images on the error rates of the algorithms are examined with practical palm print images. The result implies that considering the mean of the similarities with multiple images is useful to improve authentication accuracy.

Original languageEnglish
Title of host publicationProceedings - 4th International Conference on Intelligent Systems, Modelling and Simulation, ISMS 2013
Pages52-55
Number of pages4
DOIs
Publication statusPublished - 2013
Externally publishedYes
Event4th International Conference on Intelligent Systems, Modelling and Simulation, ISMS 2013 - Bangkok, Thailand
Duration: Jan 29 2013Jan 31 2013

Publication series

NameProceedings - International Conference on Intelligent Systems, Modelling and Simulation, ISMS
ISSN (Print)2166-0662
ISSN (Electronic)2166-0670

Conference

Conference4th International Conference on Intelligent Systems, Modelling and Simulation, ISMS 2013
CountryThailand
CityBangkok
Period1/29/131/31/13

Keywords

  • accuracy
  • authentication
  • biometrics
  • pattern matching

ASJC Scopus subject areas

  • Artificial Intelligence
  • Software
  • Modelling and Simulation
  • Theoretical Computer Science

Fingerprint Dive into the research topics of 'A note on authentication accuracy with multiple biometric images'. Together they form a unique fingerprint.

Cite this