Wavelet based method for localization of myocardial infarction using the electrocardiogram

Author

Noorian, Azadeh ; Dabanloo, Nader Jafarnia ; Parvaneh, Saman

Author_Institution

Dept. of Biomed. Eng., Islamic Azad Univ., Tehran, Iran

fYear

2014

Firstpage

645

Lastpage

648

Abstract

This paper presents detection and localization of myocardial infarction (MI) using REF neural networks classifier with wavelet coefficient as features extracted from frank leads. Detection of MI aim to classify healthy and having MI and Localization aim to specify the infracted region of the heart. The electrocardiogram (ECG) source used in the PTB database available on physio-bank. Frank lead vx, vy, vz is get from 12 lead ECG using Dower transformation. Wavelet coefficient of different levels and families of each beat are extracted. We extract wavelet coefficient in three level 3, 4, 5 from threewavelethaar, db4, db10 to evaluate the different kinds of wavelet.

Keywords

bioelectric potentials; diseases; electrocardiography; feature extraction; haemodynamics; medical signal processing; muscle; neural nets; patient diagnosis; wavelet transforms; 12 lead ECG; Dower transformation; ECG source; MI detection; MI localization; PTB database; REF neural networks classifier; electrocardiogram source; frank lead vx; frank lead vy; frank lead vz; frank lead-extracted features; heart infracted region; myocardial infarction detection; myocardial infarction localization; physio-bank; wavelet based method; wavelet coefficient extraction; Abstracts; Electrocardiography; Heart; Lead; Myocardium; Noise measurement; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

Computing in Cardiology Conference (CinC), 2014

ISSN

2325-8861

Print_ISBN

978-1-4799-4346-3

Type

conf

Filename

7043125