ANALISIS KLASIFIKASI SINYAL EKG BERBASIS WAVELET DAN JARINGAN SYARAF TIRUAN
ECG signals analysis at first associated to pattern recognition of the ECG signals marphology. Nonetheless the signals marphology varying not only in different patients but also in the same patient. The varying of the ECG marphology has efected difficulties in ECG analysis, particularly for a traini...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Theses and Dissertations NonPeerReviewed |
| Published: |
[Yogyakarta] : Universitas Gadjah Mada
2012
|
| Subjects: | |
| Online Access: | https://repository.ugm.ac.id/100689/ http://etd.ugm.ac.id/index.php?mod=penelitian_detail&sub=PenelitianDetail&act=view&typ=html&buku_id=57169 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | ECG signals analysis at first associated to pattern recognition of the ECG
signals marphology. Nonetheless the signals marphology varying not only in
different patients but also in the same patient. The varying of the ECG marphology
has efected difficulties in ECG analysis, particularly for a trainingless medicines. On
the other hand the ECG signals contain much noises. Therefore it was require the
suitable methods for ECG signals analysis. The research aim are to analyzing and
classifying the ECG signals from heart condition of normal, arrhytmia, ventricular
tachyarrhytmia, intracardiac atrial fibrillation and myocard infarction based on
wavelet transformation and artificial neural network backpropagation.
The research stages are data preparing, pre-processing, feature extraction,
processing and post-processing. The 60/50 Hz noises in ECG signals from power line
interference can be reduced using IIR notch filter with pole-zero placement method.
The baseline wander noises can be reduced using discrete wavelet transform of 11
level decomposition to find frequency component below 0,5 Hz as a noise source.
ECG feature extraction using normalization of average energy from each
decomposition signals of 6 level by symlet wavelet (sym8). Artificial neural network
backpropagation with sturcture of 7 input neuron, 7 hidden layer neuron and 5 output
neuron has been used to classification of the ECG signals.
Based on this work results obtained that average accuracy percentage of the
neural network recognized all of the ECG types reached 87,424 %. Highest accuracy
percentage of 95,455 % for ventricular tachyarrhytmia and lowest accuracy
percentage of 70 % for arrhytmia classification. |
|---|