ECG signal compression using 2-D DWT Hermite coefficients

The Electrocardiogram (ECG) signal carries very important information about the Human heart condition. Many researchers have made a significant contribution for use of optimum storage devices and improving the speed of transmission over the wired or wireless channel. These applications demand for the good and very trustworthy compression schemes. In our paper we propose a ECG compression scheme using discrete Hermite functions. The ECG signal is spread over the discrete Hermite functions basis and the 2-D wavelet based compression is carried out. The reconstruction of the signal is achieved by applying the inverse 2-D wavelet with maximal similarities in the diagnostic characters. The results of the compression scheme are checked by use of the standard performance criterion such as the compression ratio (CR), Percent root mean square difference (PRD), and normalized cross correlation coefficient (NCC). The nature of reconstructed signal is assessed by the retrieval of the dominant morphological features like P-part, QRS detection, R-part amplitude, S-part and T-part amplitudes upon reconstruction. There is a minimal loss of signal which do not exceed so that the cardiologists are not misguide for false diagnosis. The analysis is carried out by using the Haar and Deubachies(db2-db20) wavelets filters.