Acta Mathematica Academiae Paedagogicae Nyíregyháziensis, Vol. 32, No. 2, pp. 327-333 (2016)

ECG-based heart beat detection using rational functions

Zoltán Gilián

Eötvös Loránd University

Abstract: The aim of this paper is to present a novel heart beat detection algorithm using rational modelling of ECG signals. The algorithm considers several candidate beat locations. For a given candidate a rational model is fitted to the ECG signal by means of numerical optimization and Fourier partial sums with respect to the Malmquist-Takenaka system. The resultant model parameters are used as a basis of classification. The classification is performed by an SVM classifier, which is trained on annotated ECG records of the PhysioNet database.

Keywords: Rational model, Malmquist-Takenaka system, Fourier partial sum.

Classification (MSC2000): 92C55

Full text of the article:

[Previous Article] [Contents of this Number]
© 2016 FIZ Karlsruhe / Zentralblatt MATH for the EMIS Electronic Edition