Comparison of pulse active ratio similarity measurement

H Haidawati Binti Mohamad Nasir, Sairul Izwan Bin Safie, K Kushsairy Bin Abdul Kadir, John Soraghan, Lykourgos Petropoulakis

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

Abstract

Distance measurement is a quantitative tool to measure similarity or dissimilarity between two objects. A correct selection of distance measurement will enhance the performance of a biometric authentication system. In this paper, various types of distance measurement methods such as Euclidean distance, City block distance, Chebyshevdistance, Minkowskidistance, Sorensen distance, Cosine distance and Mahalanobis distance are evaluated to determine the best similarity measure to be used with the novel pulse active ratio (PAR) feature extraction method. The results are obtained based on comparing 486 electrocardiography (ECG)signals which provide a total of 42,149 ECG comparisons. The comparisons show that the similarity measurement based on Sorensen distance gives the best matching algorithm to increase the performance of the PAR ECG biometric approach.
Original languageEnglish
Title of host publication1st International Conference on Artificial Intelligence, Modelling and Simulation (AIMS), 2013
PublisherIEEE
Pages309 - 314
Number of pages6
ISBN (Print)978-1-4799-3250-4
DOIs
Publication statusPublished - 5 Dec 2013

Keywords

  • distance measurement
  • electrocardiography
  • feature extraction
  • pulse measurements

Fingerprint

Dive into the research topics of 'Comparison of pulse active ratio similarity measurement'. Together they form a unique fingerprint.

Cite this