Feasibility of neural network based QRS-T cancellation schemes for P-wave detection

Author

Vásquez, C. ; Hernandez, Alfredo I. ; Carrault, G. ; Mora, F.A. ; Passariello, G.

Author_Institution

Grupo de Bioingenieria y Biofisica Aplicada, Univ. Simon Bolivar, Caracas, Venezuela

fYear

1998

fDate

13-16 Sep 1998

Firstpage

625

Lastpage

628

Abstract

This work reviews the possible application of artificial neural networks (ANNs) to the problem of ventricular activity (VA) cancellation. The system proposed consists of estimating a non-linear time-varying transfer function between two ECG channels using a Time Delay Neural Network (TDNN). Three different TDNN topologies (purely feed-forward, simple recurrent, and fully recurrent) were implemented and tested using record 108 of the MIT-BIH DB. In order to evaluate these networks, two performance measures were introduced in this work, namely VA energy change and Signal to Noise Ratio (SNR) improvement. Preliminary results using VA energy change and SNR improvement indicators showed that the simple recurrent topology presents the best performance

Keywords

electrocardiography; feedforward neural nets; medical signal detection; medical signal processing; recurrent neural nets; ECG channels; P-wave detection; artificial neural networks; electrodiagnostics; fully recurrent neural nets; neural network based QRS-T cancellation schemes; nonlinear time-varying transfer function; purely feedforward neural nets; simple recurrent neural nets; time delay neural network; Artificial neural networks; Delay effects; Delay estimation; Electrocardiography; Feedforward systems; Network topology; Neural networks; Signal to noise ratio; Time varying systems; Transfer functions;

fLanguage

English

Publisher

ieee

Conference_Titel

Computers in Cardiology 1998

Conference_Location

Cleveland, OH

ISSN

0276-6547

Print_ISBN

0-7803-5200-9

Type

conf

DOI

10.1109/CIC.1998.731951

Filename

731951