Combine crossing matching scores with conventional matching scores for bimodal biometrics and face and palmprint recognition experiments

Yong Xu, Qi Zhu, Dapeng Zhang

Research output: Journal article publicationJournal articleAcademic researchpeer-review

45 Citations (Scopus)

Abstract

In this paper, we define for the first time the crossing matching score of two biometrics traits and combine it with the conventional matching scores to perform personal authentication. The proposed method is very suitable for the bimodal biometrics systems with two similar biometrics traits such as the system with visible light and infrared face images and the system with palm images captured at two bands. The proposed method first runs for the first and biometrics traits, respectively. For each of these two biometrics traits, the proposed method calculates the matching scores between the testing sample and each training sample. The matching scores generated from the first and second traits are referred to as the first and second matching scores, respectively. Second, the proposed method calculates the crossing matching scores, i.e. the matching scores between the testing sample of the second biometrics trait and the training samples of the first biometrics trait. Finally, we use a weighted fusion scheme to combine the first, second and crossing matching scores for personal authentication.
Original languageEnglish
Pages (from-to)3946-3952
Number of pages7
JournalNeurocomputing
Volume74
Issue number18
DOIs
Publication statusPublished - 1 Nov 2011

Keywords

  • Biometrics
  • Matching score level fusion
  • Pattern recognition

ASJC Scopus subject areas

  • Computer Science Applications
  • Cognitive Neuroscience
  • Artificial Intelligence

Cite this