ECG Signal De-Noising and Compression Using Discrete Wavelet Transform and Empirical Mode Decomposition Techniques

Electrocardiogram (ECG) consists of recording electrical activity of the heart over time and recognized biological signal used for clinical diagnosis .The ECG signal is very sensitive in nature, and if small noise is embedded or mixed with original signal during acquisition, the various characteristics of the signal change, thus their analysis becomes difficult. In this paper, ECG signal can be roughly divided into two by functionality: preprocessing and feature compress. The preprocessing stage removes or suppresses noise and compresses the data for efficient storage or transmission. In the first part of this paper, two de-noising techniques for ECG signals based on Discrete Wavelet Transform (DWT) and Empirical Mode Decomposition (EMD) are proposed. The proposed methods are evaluated based on the Signal to Noise Ratio (SNR), Mean Square Error (MSE) and Percent Root mean square Difference (PRD) where white Gaussian noise is artificially added with original signal. The result of the comparison shows that the DWT technique is performance. In the second part of this paper, the DWT technique is used for compression. ECG signal compression is necessary for efficient storage or transmission. Discrete Wavelet Transform is a recently developed compression technique in signal compression. DWT signal compression includes decomposition (transform of signal with an appropriate wavelet family); hard thresholding is applied to the detail coefficients and compute wavelet reconstruction using the original approximation coefficients and the modified detail coefficients. Then the result is obtained by applying technique of local thresholding. The optimal value of the threshold (T) is determined such that the reconstructed signal is close to the original one as much as possible such that the maximum percent of Retained Energy in the compressed signal is obtained. The measure used to evaluate the quality of the compressed signal is the measure of PRD and modified PRD (MPRD). The de-noising and compression algorithms was implemented and tested upon records selected from the MIT – BIH arrhythmia database