基于Levenberg-Marquardt算法的心电高斯模型研究

Levenberg-Marquardt Algorithm-Based Gaussian Model Research for ECG

针对心电信号中各构成成分(P波、QRS波、T波)波形特征,本研究提出了一种基于伯格–马夸尔特算法创建心电高斯模型以实现心电波形特征研究进而对心电特征进行定性评价研究。首先,基于小波分解对心电信号进行预处理,以消除基线漂移以及背景噪声;其次,以心电信号各R峰间距为参考,按照1:2的分割比例对心电信号进行周期分割;最后,基于Levenberg-Marquardt (LM)算法创建心电高斯混合模型以刻画各成分波形特征。本算法通过对MIT-BIH心律失常数据库中48条心电数据进行分析,利用其拟合优度均值为0.9668对本算法可行性进行了验证。

Aimed at the waveform characteristics of main components (P wave, QRS complex and T wave) to an electrocardiogram (ECG), this study proposes a novel method based on Levenberg-Marquardt algorithm (LMA) to build the Gaussian model for researching ECG signal, and then to character the features of each component. Firstly, the wavelet decomposition-based preprocessing method is employed to cancel baseline shift and background noise existing in ECG signal. Secondly, ECG signal is segmented into periodic signal based on the distance between tow sequence R peaks as well its ratio of 1:2. Finally, LMA-based Gaussian Mixture model is built to character each component waveform. The performance of the proposed method has been evaluated by using the 48 ECG records of MIT-BIH arrhythmia database, and the higher average of R-square value 0.9668, used to measure the goodness of fit, has been achieved.