OPTIMAL ECG DATA COMPRESSION FOR AN e-HEALTH TRANSMISSION SYSTEM

Abstract

In the present work optimal compression of ECG has been carried out. Transform based methodology viz Discrete Cosine Transform (DCT), Discrete Wavelet Transform (DWT). and Fast Fourier Transform (FFT) methods are used for compression. Prior to ECG data compression and after data retrieval features of ECG signal like QRS complex, heart rate, P wave, T wave and PR interval have been extracted by Wavelet Transform Algorithm and Pan-Tompkins criteria. The extracted features of reconstructed ECG signals are compared with those of the original ECG signal data. In order to evaluate the performance of the technique the algorithm has been applied to twenty normal records of the MlT-BIH database each of one minute duration. The performance of compression is evaluated in terms of Compression Ratio (CR), Percentage Root Mean Square Difference (PRD), Signal to Noise Ratio (SNR), and Quality of Score (QS). PRD for compressed signal is kept under 7% to ensure that diagnostic information is retained; the magnitude of PRD being chosen after exhaustive analysis of ECG data. Best results were obtained by DCT method the value of CR being 15:1 and of PRD 5%.