A Variety of Electrocardiogram (ECG) Signal Processing

Authors

  • Ali Nazem Kamber , Alyaa Mohammed jawad , Mohammed Ameer Altamimi , Hussein Alaa Alkaabi

Abstract

Intelligent systems are increasingly applied to engineering sciences, especially to diagnose multiple diseases.
The present paper will address a new method of ECG signal processing. In this plan, in order to achieve a high
rate of classification accuracy as well as accurate diagnosis of the disease, the importance of selecting the best
methods of feature extraction and also selecting an appropriate classification method is discussed. In general,
these processes can be divided on the ECG signal into three stages: preprocessing, feature extraction and
classification. In the first stage, by eliminating all kinds of noise, the signal quality is improved, and in the next
stage, important signal properties are extracted. Finally, previously derived features were reduced by pca
algorithm that led to an increase in the system’s detection rate. Then, arrhythmia recognition was performed via
neural networks and support vector machine, which ultimately amounted accuracy to 99.3 and 99.88 for
classifying neural networks and support vector machine, respectively

Published

2020-12-03

Issue

Section

Articles