Efficient ECG data compression and transmission algorithm for telemedicine

Abstract

This paper presents an efficient electrocardiogram (ECG) data compression and transmission algorithm based on discrete wavelet transform and run length encoding. The proposed algorithm provides comparatively high compression ratio and low percent root-mean-square difference values. 48 records of ECG signals are taken from MIT-BIH arrhythmia database for performance evaluation of the proposed algorithm. Each record of ECG signals are of duration one minute and sampled at sampling frequency of 360 Hz over 11-bit resolution. Discrete wavelet transform has been used by means of linear orthogonal transformation of original signal. Using discrete wavelet transform, signal can be analyzed in time and frequency domain both. It also preserves the local features of the signal very well. After thresholding and quantization of wavelet transform coefficients, signals are encoded using run length encoding which improves compression significantly. The proposed algorithm offers average values of compression ratio, percentage root mean square difference, normalized percentage root mean square difference, quality score and signal to noise ratio of 44.0, 0.36, 5.87, 143, 3.53 and 59.52 respectively over 48 records of ECG data.