RANDOM INTERPOLATION AVERAGE FOR ECG SIGNAL DENOISING USING MULTIPLE WAVELET BASES
Ying Yang, Yusen Wei · Biomedical Engineering Applications Basis and Communications · 2013
The random interpolation average (RIA) is a simple yet good denoising method. It firstly employed several times of random interpolations to a noisy signal, then applied the wavelet transform (WT) denoising to each interpolated signal and averaged all of the denoised signals to finish the denoising process. In this paper, multiple wavelet bases and the level-dependent threshold estimator were used in the RIA scheme so that it can be more suitable for the electrocardiogram (ECG) signal denoising. The synthetic ECG signal, real ECG signal and four types of noise were used to perform comparison experiments. The results show that the proposed method can provide the best signal to noise ratio (SNR) improvement in the deoising applications of the synthetic ECG signal and the real ECG signals. For the real ECG signals denoising, the average SNR improvement is 5.886 dB, while the result of the RIA scheme with single wavelet basis (RIAS), the fully translation-invariant [TI (fully)] and the WT denoising using hard thresholding [WT (hard)] are 5.577, 5.274 and 3.484 dB, respectively.