Touchscreen swipe based biometric authentication using spectral regression - Kernel discriminant analysis Spektral regresyon - Çekirdek diskriminant analizi kullanarak dokunmatik ekran kaydirmaya dayali biyometrik kimlik doǧrulama


SİVAZ O., AYKUT M.

29th IEEE Conference on Signal Processing and Communications Applications, SIU 2021, Virtual, Istanbul, Turkey, 9 - 11 June 2021 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/siu53274.2021.9477815
  • City: Virtual, Istanbul
  • Country: Turkey
  • Keywords: swipe biometric, sr-kda, ls-svm
  • Karadeniz Technical University Affiliated: Yes

Abstract

© 2021 IEEE.Nowadays, various biometric verification systems are used to ensure information security. In our study, the verification of the individuals from the swipe gestures on the touch screen is discussed. In this context, firstly, feature vector was created from the swipe gestures. Spektral Regression-Kernel Discriminant Analysis algorithm which casts discriminant analysis into a regression framework by using spectral graph analysis was aplied to the feature vector and then classified with Least Squares Support Vector Machines. For the experiments, publicly available JSS18 Database, which was constituted by the up, down, left and right swipe gesture of 31 individuals, acquired horizontally and vertically in eight sessions, was used and performance evaluation has been performed through EER. When the results are analyzed, it can be seen that the proposed approach increased the success and a significant reduction in error has been achieved when compared to the methods in the literature.