基于三维统计距离的异常心电节律检测方法

本文提出了一种异常心电节律的检测方法.首先重构出心电时间序列的二维相平面,并在此相平面上加入时间轴形成一个封闭的三维空间,通过比较时间序列的重构三维向量序列在此三维空间中的概率分布,得到两时间序列的差异,从而实现异常节律的检测.本方法不但能够检测出两心电信号在波形上存在的差异,并且可以检测出它们之间存在的各波(P、QRS和T波)位置的差异.本文对此方法和基于二维相平面的二维统计距离方法进行了异常节律检测的实验比较,结果表明此方法具有更好的检测效果.

Abnormal Rhythm Detection Based on 3-Dimensional Statistical Distance

A new method is proposed to detect abnormal rhythm in this paper. First, we reconstruct a 2-dimensional phase space from ECG time series and form a closed 3-dimensional space by adding a time axis. Then we get the 3-dimensional vector series and their probability distribution on the 3-dimensional space. Finally, we calculate the statistical distance from probability distributions and use it to detect abnormal rhythm. Our method could detect not only morphological changes, but also changes of locations of waveforms. By comparing to the 2-dimensional statistical distance method, we conclude that our method has a better performance.