ANALYSIS OF LINEAR QUANTIZATION BASED WAVELET DECOMPOSITION TECHNIQUE FOR ELECTROCARDIOGRAM SIGNAL COMPRESSION

Efficient ECG signal compression technique is presented in this paper based on the wavelet transformation. Decomposition is achieved by tree structure of wavelet filters with desired decomposition level performance. The proposed compression technique is exposed to ECG signals which are taken from MIT-BIHA database of 48 ECG records for test and validation. Method is analyzed and compared on the parameters of compression ratio (CR), percentage root mean square difference (PRD%), signal to noise ratio (SNR) with optimum level of decomposition. Proposed technique achieves CR 35.77:1, PRD% 0.991 and SNR 40.56dB with decomposition level of 8 and 8bits/ sample quantizing value. The result shows significant improvement in parametric values and achieves good balance between compression efficiency and its reconstruction quality while keeping low compression error.