Premium
ECG biometric recognition using SVM‐based approach
Author(s) -
Rezgui Dhouha,
Lachiri Zied
Publication year - 2016
Publication title -
ieej transactions on electrical and electronic engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.254
H-Index - 30
eISSN - 1931-4981
pISSN - 1931-4973
DOI - 10.1002/tee.22241
Subject(s) - biometrics , heartbeat , support vector machine , pattern recognition (psychology) , artificial intelligence , hyperparameter , computer science , normal sinus rhythm , classifier (uml) , speech recognition , atrial fibrillation , medicine , computer security , cardiology
This paper presents a new approach for biometric personal identification based on electrocardiogram (ECG) features. ECG, which reflects cardiac electrical activity, is a distinctive characteristic of a person and can be used for security needs. Twenty‐one features based on temporal and amplitude distances between detected fiducial points and 10 morphological descriptors are extracted from each heartbeat. Then, support vector machine (SVM) is used as a classifier. A comparative study between two kernels, Gaussian and polynomial, was made in order to determine the best kernel and the appropriate values of hyperparameters that improve the recognition performance. The algorithm is evaluated using two databases, namely MIT‐BIH Arrhythmia and MIT‐BIH Normal Sinus Rhythm. Analysis of the results shows that the combination of all features allows improvement of our system efficiency with regard to healthy human subjects and those with arrhythmia. © 2016 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.