Summary
International Technical Conference on Circuits/Systems, Computers and Communications
2016
Session Number:W2-1
Session:
Number:W2-1-2
Deriving ECG from EASI Electrodes via Machine Learning
Piroon Kaewfoongrungsi, Daranee Hormdee ,
pp.783-786
Publication Date:2016/7/10
Online ISSN:2188-5079
DOI:10.34385/proc.61.W2-1-2
PDF download (1.5MB)
Summary:
The 12-lead Electrocardiogram (ECG) is the standard clinical method of heart disease diagnose. Measuring all 12 leads is often impractical. In 1988, Gordon Dower has introduced an EASI-lead ECG System. In order to gain all 12-lead ECG back from this EASI-lead system, Dower's equation was proposed then. Ever since various attempts have been explored to improve the synthesis accuracy, mostly via linear regression. This paper presents how machine learning was used to find a set of transfer function for deriving the 12-lead ECG from EASI-lead system. The experiments were conducted to compare the results those of Support Vector Regression (SVR), Artificial Neural Networks (ANNs) against those of Dower's method. The results have shown that the best performance amongst those methods with the less RMSE error values for all signals with the standard 12-lead ECG was obtained by SVR, followed ANNs and Dower's equation, respectively.