Denoising method of ECG signal with power threshold function under wavelet transform and smoothing filter✱

Abstract

An electrocardiogram (ECG) is an important instrument for doctors to diagnose heart diseases. It is an electrical signal that evolves from the heart and changes over time. It is vulnerable to interference from various low-frequency and high-frequency noise. This paper proposes a new adaptive power threshold function to achieve the denoising of ECG signals. On the basis of wavelet transformation and smooth decomposition, the power threshold function is used to perform an adaptive threshold denoising on the decomposed signal with high frequency noise. The signal is reconstructed from the denoised high frequency components and useful components and coefficients of the remaining layers are set at zero. Taking the ECG signal in the MIT-BIH ECG database as the original data, adding different degrees of Gaussian white noise for experimental analysis, it is proved from the quantitative and qualitative aspects that the proposed method has superiority in removing the noise of the ECG signal compared with the traditional threshold function.