ECG Signal Compression using Normalization and Thresholding

ABSTRACT:        

In this paper, ECG signal compression using normalization and thresholding of discrete cosine transform (DCT) and Wavelet Transform (WT) signal coefficients is studied. First ECG signal is transformed using Linear Transforms (DCT and WT). The transformed coefficients (TC) are then normalized by dividing the TC by  (where, is number of samples). To reduce the range of TC, the normalized coefficients (NC) are thresholded by bisection algorithm. The thresholding is done in a way so that error between actual PRD (percent root mean square difference) and UPRD (user specified percent root mean square difference) remains within the specified limits. The binary look up table is made to store the position of zeroed and non-zeroed coefficients and this table is encoded by Huffman coding. The non-zero thresholded coefficients are then quantized and encoded by arithmetic coding.  The results presented in this paper are compared with [9], where the same ECG compression techniques used without normalization. The comparison indicates that the normalization of DCT and WT coefficient before thresholding improves the performance. The results are presented on different ECG signals of varying characteristic.

Keywords: Discrete cosine transform, ECG Signal, quality controlled compression, Wavelet Transform, Normalization

http://www.ijbst.org/Home/papers-published/ijbst-2012-volume-5-issue-8