ECG Based Expert System for Helping Doctors to Diagnosis Heart Abnormalities

Authors

  • Jannat Ul Maowa STUDENT
  • Mohammad Osiur Rahman
  • Syed Muhammad Asifur Rahman

Keywords:

ECG signal, Cardiac cycle, Feature extraction, P-QRS-T segment

Abstract

Electrocardiogram (ECG) is one of the most significant investigative tools for the identification of any abnormalities in heart function. The correct ECG interpretation is essentially required in order to evaluate the valuable information inside the ECG signal. The conventional technique of pictorial analysis to inspect the ECG signals by doctors or physicians are not effective as well as time consuming. This process also requires large storage space and extensive manual effort. In this research, a computer-aided application model for classification of ECG signals for detection of cardiac abnormalities is discussed. The model is based on some prevailing procedures in literature which were adapted to outfit our application. The developed model comprises ECG signal pre-processing, extraction of various morphological features using wavelet transform and classifying it with support vector machine. In this paper, for making the application (a software tool) user friendly, we present the development of a MATLAB based graphical user interface (GUI) for the application, which then makes it serve completely as a cost-effective computer aided application to evaluate ECG signals for detection of heart irregularities. The performance of the developed application was inspected using 120 sample of ECG image. Accuracy of 95%, sensitivity of 95.23%, specificity of 94.7%, positive predictivity of 95.23% and detection error rate of 9.5% displays that the developed application can be employed for the identification of heart diseases in patients. Also, the application is fast, easy to use and gives the result of diagnosis as ‘Normal’ or ‘Abnormal’.

Electrocardiogram (ECG) is one of the most significant investigative tools for the identification of any abnormalities in heart function. The correct ECG interpretation is essentially required in order to evaluate the valuable information inside the ECG signal. The conventional technique of pictorial analysis to inspect the ECG signals by doctors or physicians are not effective as well as time consuming. This process also requires large storage space and extensive manual effort. In this research, a computer-aided application model for classification of ECG signals for detection of cardiac abnormalities is discussed. The model is based on some prevailing procedures in literature which were adapted to outfit our application. The developed model comprises ECG signal pre-processing, extraction of various morphological features using wavelet transform and classifying it with support vector machine. In this paper, for making the application (a software tool) user friendly, we present the development of a MATLAB based graphical user interface (GUI) for the application, which then makes it serve completely as a cost-effective computer aided application to evaluate ECG signals for detection of heart irregularities. The performance of the developed application was inspected using 120 sample of ECG image. Accuracy of 95%, sensitivity of 95.23%, specificity of 94.7%, positive predictivity of 95.23% and detection error rate of 9.5% displays that the developed application can be employed for the identification of heart diseases in patients. Also, the application is fast, easy to use and gives the result of diagnosis as ‘Normal’ or ‘Abnormal’.

Downloads

Published

2020-12-31

How to Cite

Maowa, J. U., Rahman, M. O., & Rahman, S. M. A. (2020). ECG Based Expert System for Helping Doctors to Diagnosis Heart Abnormalities. ournal of nergy and nvironment, 12(2). etrieved from http://journal.uniten.edu.my/index.php/jee/article/view/189

Issue

Section

Normal Submission