The comparison between linear regression derivings of 12-lead ECG signals from 5-lead system and EASI-lead system

摘要

The measuring for diagnosis of heart disease typically use 12-lead Electrocardiogram (ECG) monitoring system with 10 electrodes, but quantifying all 12 leads is often cumbersome and impractical especially on a long term monitoring. Various reduced number of lead systems have been proved to work in the past. A number of research have been explored on deriving the missing signals from these fewer measured leads. Two well-known aspects of reduced electrodes ECG systems are either using subset of the 12-lead system or using the complete new lead scheme called EASI-lead system. This paper presents the comparison between these 2 fewer-electrode approaches, 5-lead system with V1 and EASI-lead system. Due to its simplicity, Linear Regression is the choice for derivation along with 5-fold Cross-validation on data set from PhysioNet Database. Surprisingly, despite half of the signals can be measured straightforward in the 5-lead system with V1, the experimental results show that the better RMSEs come from the EASI-lead system.