Performance of Wavelet Filters for ECG Compression Based on Linear Predictive Coding Using Different Thresholding Functions

Abstract

In this paper, a linear predictive coding (LPC) based electrocardiogram (ECG) signal compression technique is examined and improved by exploiting different wavelet filters and thresholding functions. In this method, ECG signals are captured from MITBIH arrhythmia database and compression is achieved by propagating wavelet coefficients of ECG signal to linear predictive coder. Then different thresholding techniques such as hard, soft, hybrid, soft soft and firm thresholding are used. At receiver end, reverse procedure of this method is applied and original ECG signal is reconstructed. The simulation results show that the wavelet filter 'db4' gives improved results as compared to other wavelet filters. Firm thresholding performs better than other thresholding techniques in terms of fidelity parameters such as signal to noise ratio (SNR), percentage root mean square difference (PRD), maximum error (ME) and compression ratio (CR).