Digital Signal Processing 17 (2007) 319–326 www.elsevier.com/locate/dsp ECG compression method using Lorentzian functions model Abdelaziz Ouamri a , Amine Naït-Ali b,∗ a Université des Sciences et de la Technologie d’Oran, Département Electronique, Laboratoire LSI, BP 1505, Oran El’Mnouar, 31 Oran, Algeria b Université Paris 12, LISSI EA 3956, 61 avenue du Général de Gaulle, 94010 Créteil, France Available online 8 August 2006 Abstract An ECG compression algorithm using a combination of Lorentzian functions model is proposed in this paper. In order to estimate the parameters of the Lorentzian functions, the discrete Fourier transform (DFT) is first applied to a mean removed ECG signal from which only the most significant DFT coefficients are retained. The obtained coefficients are, then modeled as the sum of a given number of superimposed exponentially damped sinusoids (EDS), commonly identified by their amplitudes, real damping factors, frequencies and initial phases. Finally, these EDS parameters are estimated, using SVD method, then coded. The algorithm has been tested for its coding efficiency and reconstruction capability by applying it to several popular, benchmark ECG signals. Encouraging results have been obtained.  2006 Elsevier Inc. All rights reserved. Keywords: ECG compression; SVD; Lorentzian functions; Exponentially damped sinusoids 1. Introduction It is well known that the storage and transmission of digital ECG signals are of great importance for various applications. In these applications, compressing ECG waveforms without any considerable degradation in quality is of a great consideration. The ECG compression techniques can be classified in three broad categories: direct methods, transform-based methods, and parameter extraction methods. In the direct methods category, the original samples are directly compressed [1,2]. In the transformation methods category, the original samples are transformed and the compression is performed in the new domain. Among, the algorithms which employ transform-based techniques, there are several algorithms based on discrete cosine transform [3], and wavelet transform [4–7]. In the category of the methods with extraction of parameters, one extracts some features of the signal that are used later for its reconstruction [8,9]. This paper presents an algorithm based on modeling the ECG signal as a linear combination of Lorentzian func- tions, to achieve effective ECG data compression. The estimation of unknown parameters of Lorentzian functions from a finite number of noisy data observations has been considered in many applications, particularly in NMR biomedical signals analysis [10]. * Corresponding author. Fax: +33 1 45 17 14 92. E-mail address: naitali@univ-paris12.fr (A. Naït-Ali). 1051-2004/$ – see front matter  2006 Elsevier Inc. All rights reserved. doi:10.1016/j.dsp.2006.07.003