DATA COMPRESSION OF ECG USING DWT

Abstract

The proposed compression scheme is based on the Energy packing Efficiency (EPE %) of the signal. A threshold value is calculated based on the EPE(%) calculated of the wavelet coefficients. Data taken from the MIT-BIH database is analysed with the help of the Wavelet Transform. Wavelet Transform divide the signal mainly into two type of bands, approximation coefficients band and detail coefficients bands. In the proposed scheme, a threshold value is calculated for the all detail sub-bands based on the EPE(%) calculated. No threshold value is calculated for the approximation band coefficients. EPE(%), Compression ratio (CR) and Percentage Root Mean Square Difference (PRD %) is also calculated of the reconstructed signal to check the signal quality and compression performance. Calculation of the EPE(%) of the reconstructed signal is used in terms of retained energy or information of the signal. If the signal quality is poor then again calculate the threshold value for each detail sub-band with help of the EPE(%) and check signal quality. In the proposed scheme, different wavelet transforms (Bi-orthogonal, orthogonal, coiflect) are used for the comparison purpose. Different data sets are analysed using different wavelets and analyse the compression performance of these wavelet families using performance parameters used. The effect of decomposition level and order of the wavelets is also analysed.