ECG De-noising Based on Empirical Mode Decomposition

Electrocardiogram (ECG) signal is useful in diagnosing the heart condition. Good quality ECG is utilized by physicians for interpretation and identification of physiological and pathological phenomena. However, The electrocardiogram (ECG) signal may mix various kinds of noises while gathering and recording. In this paper, we propose a new ECG enhancement method based on the recently developed empirical mode decomposition (EMD). The proposed EMD-based method is able to remove noise from the munder a wide range of variations for noise. The method is validated through experiments on the MIT-BIH databases. The simulations show that that the proposed methods in the paper provide better performance of noise reduction than wavelet thresholding de-noising methods in aspects of remaining geometrical characteristics of ECG signal and the signal-to-noise ratio (SNR).