Charles Darwin University

CDU eSpace
Institutional Repository

 
CDU Staff and Student only
 

Improved ECG signal analysis using wavelet and feature extraction

Matsuyama, Aya, Jonkman, Mirjam E. and De Boer, Friso G. (2007). Improved ECG signal analysis using wavelet and feature extraction. Methods of Information in Medicine: journal of methodology in medical research, information and documentation,46(2):227-230.

Document type: Journal Article
Citation counts: Google Scholar Search Google Scholar

IRMA ID 77514990xPUB15
Title Improved ECG signal analysis using wavelet and feature extraction
Author Matsuyama, Aya
Jonkman, Mirjam E.
De Boer, Friso G.
Journal Name Methods of Information in Medicine: journal of methodology in medical research, information and documentation
Publication Date 2007
Volume Number 46
Issue Number 2
ISSN 0026-1270   (check CDU catalogue  open catalogue search in new window)
Start Page 227
End Page 230
Total Pages 4
Place of Publication Stuttgart, Germany
Publisher Schattauer
Field of Research INFORMATION AND COMPUTING SCIENCES
1117 - Public Health and Health Services
0903 - Biomedical Engineering
HERDC Category C1 - Journal Article (DEST)
Abstract Automatic detection of arrhythmias is important for diagnosis of heart problems. However, in ECG signals, there is significant variation of waveforms in both normal and abnormal beats. It is this phenomenon, which makes it difficult to analyse ECG signals. The aim of developing methodology is to distinguish between normal beats and abnormal beats in an ECG signal. METHODS: ECG signals were first decomposed using wavelet transform. The feature vectors were then extracted from these decomposed signals as normalised energy and entropy. To improve the classification of the feature vectors of normal and abnormal beats, the normal beats which occur before and after the abnormal beats were eliminated from the group of normal beats. RESULTS: With our proposed methods, the normal beats and abnormal beats formed different clusters of vector points. By eliminating normal beats which occur before and after the abnormal beats, the clusters of different types of beats showed more apparent separation. CONCLUSIONS: The combination of wavelet decomposition and the classification using feature vectors of the beats in ECG signals separate abnormal beats from normal beats. The elimination of the normal beats which occur before and after the abnormal beats succeeded in minimising the size of normal beats cluster.
 
Versions
Version Filter Type
Access Statistics: 77 Abstract Views  -  Detailed Statistics
Created: Fri, 12 Sep 2008, 08:35:25 CST by Administrator