ECG comptression using run length encoding

A new approach of run length encoding (RLE) is proposed in this research to compress discrete cosine transform (DCT) coefficients of time domain ECG signals. Energy compaction property of DCT facilitates the process of length encoding by accumulating the correlative coefficients into separate segments. Thus the high probability of redundancies in consecutive coefficients facilitates the use of RLE. To increase the CR, two stages of RLE are performed on the quantized DCT coefficients. Then binary equivalent of RLE values are obtained by applying Huffman coding. Finally the distortion indices of relevant clinical diagnostic information of the reconstructed signal is measured in terms of weighted diagnostic distortion (WDD), percentage root-mean-squared difference (PRD) and root-mean-square (RMS) error indices. Results indicate that for MIT-BIH Arrhythmia database Record 117, the proposed compression algorithm can achieve a compression ratio of 14.87 with a bit rate of 185 bps.