基于灰度信息度量的阵发性房颤自动检测方法

Automatic Paroxysmal Atrial Fibrillation Detection Method Based on Grey Information Measurement

针对心房颤动这类常见的心律失常疾病,提出了一种基于灰度信息度量的阵发性房颤自动检测方法。首先,采用离散小波变换将原始心电信号进行分解;其次,选择恰当的频率子带信号并对其小波系数进行差分运算,得到一阶中心差分散点图以及对应的灰度直方图;最终,分别计算灰度方差、灰度变异系数及香农熵,作为房颤心电的融合特征。将所提取的融合特征结合超限学习机,完成了阵发性房颤的自动检测。采用MIT-BIH数据库中的数值进行实验,结果表明,所提方法能够快速有效地完成房颤心电的识别,在交叉检验数值实验结果中,准确率、敏感度、特异度分别平均达到94.0%、94.6%、93.7%。

This paper proposes an automatic paroxysmal atrial fibrillation（PAF）detection method based on grey information measurement.Firstly,the discrete wavelet transform（DWT）is applied to decompose an electrocardiogram（ECG）signal into sub-band signals.Then,the difference operation is used for wavelet coefficients to obtain the corresponding one-order central difference plot and gray histogram.Next,the variance,coefficient of variation,and Shannon entropy are extracted from the gray histogram to be the fusion features of atrial fibrillation ECG.Finally,PAF detection is completed automatically by integrating the extracted features with extreme learning machine（ELM）.Experimental results on MIT-BIH database show that the average sensitivity,specificity and accuracy of the proposed method reach 93.7%,94.6% and 94.0%,respectively.