ECG arrhythmias classification using wavelet transform and neural networks

  • Authors:

  • A. R. Sahab;Y. Mehrzad Gilmalek

  • Affiliations:

  • Electrical Group, Engineering Faculty of Islamic Azad University, Iran;Islamic Azad University Roudsar, Roudsar, Iran

  • Venue:
  • MMES'10 Proceedings of the 2010 international conference on Mathematical models for engineering science
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this research, a new method for heart Arrhythmias classification based on Wavelet Transform and Neural Networks has been proposed. Discrete Wavelet Transform (DWT) is normally used for processing and extracting the time and frequency characteristics (specifications) of ECG records. In this work, the obtained features from Wavelet Transform and morphological features of the ECG is combined with time of features this signal in order to use its results as final features to teach and test a Multi Layer Perceptron (MLP) Neural Network. In this research, 189 heart signal samples existed in MIT-BIH data base are utilized in order to teach and test the classifier. The best accuracy of 97.33 percent have been achieved for three different class of ECG signals including; Normal rhythm RBBB and LBBB.