基于抗混叠小波变换的胎儿心电信号分离方法

Fetal ECG separation method based on anti-aliasing wavelet transform

目的胎儿心电图能够较好地反映胎儿在子宫内的发育状况,但是由于采集的胎儿心电信号中混有噪声干扰,给医学诊断带来极大干扰。抗混叠小波变换算法能够从混有噪声干扰的源信号中提取胎儿心电信号,且当胎儿心电信号与母体心电信号混叠时,该方法仍能够提取胎儿心电信号。基于此,本文提出一种基于抗混叠小波变换的胎儿心电信号分离方法。方法首先对原始心电信号进行滤波预处理,再利用小波变换分离母体心电信号和胎儿心电信号,最后根据抗混叠分离算法获取混合心电信号中的胎儿心电信号,得到满周期的胎儿心电信号。结果该方法能够较好地获取胎儿心电波形,胎儿心电波形识别准确率可达100%,在信噪比较低的情况下,识别准确率仍可达到77.78%。应用此算法在国外MIT-BIT心电信号数据和国内医院临床心电信号数据中进行实验仿真,并与先前学者的胎儿心电信号提取方法进行对比。结论此方法具有较高的识别准确率以及在临床应用中的可靠性和可行性。

Objective Fetal electrocardiogram can reflect the development of fetus in the womb better,but the interference of noise in the collected fetal ECG signals brings great interference to the medical diagnosis.The anti-aliasing wavelet transform algorithm can extract the fetal ECG signals from the source signals with noise interference,and when the fetal ECG signals are aligned with the maternal ECG signals,the method can still extract the fetal ECG signals.And on account of which,this paper proposes a method of fetal ECG separation based on anti-aliasing wavelet transform.Methods First,the ECG signals are processed by noise filtering.Then,we use wavelet transform to separate maternal ECG and fetal ECG.Finally,the fetal ECG signals in the mixed ECG are obtained by using the anti-aliasing separation algorithm.Results This method can obtain the ECG waveform better,and the accuracy of fetal ECG signal recognition can be 100%.In the case of low signal-to-noise ratio,the accuracy of fetal ECG signal recognition is still 77.78%.The algorithm is simulated in foreign MIT-BIT ECG database and clinical ECG data in domestic hospitals and is compared with the previous methods of fetal ECG extraction.Conclusions The high recognition accuracy of this method,as well as its reliability and feasibility in clinical application is proved.