Segment-Specific EASI Coefficients for Improving Accuracy of Derived 12-Lead Electrocardiography

Abstract

In order to reduce the number of electrodes in standard 12-lead electrocardiography (ECG) measurement, a number of research works have been conducted. One of the attempts was deriving 12-lead ECG from the EASI lead system, which was initiated by Dower. The previous techniques, including Dower with an improved coefficient, linear regression (LR), and polynomial least square (PLR), calculated the EASI coefficients for one cycle ECG signal, i.e., covering the entire segments. To increase the accuracy, we proposed an approach that calculates the EASI coefficients by segmenting the ECG signal into three segments: PR interval, QRS complex, and ST interval. In other words, coefficients are specific for each segment. Root mean squared error (RMSE) was used to measure the performance. The results showed that our proposed approach outperformed the conventional EASI coefficient calculation in Dower's method and LR, as well as $2^{\text{rd}_{-}}$, and $3^{\text{rd}}$ -degree PLR.