ECG Signal Characterization Using Lagrange-Chebyshev Polynomials

Abstract

An ECG (electrocardiogram) is a signal representing a combination of potentials that reflect the electrical activity of the heart. These signals are often corrupted by artifacts of high magnitude and frequency during acquisition, storage and transmission. Addition of these artifacts may change the morphology of the ECG signals and can hinder accurate interpretations; hence proper characterization is required for better clinical evaluation. In this paper ECG signals from MIT-BIH database are characterized using polynomial interpolation method. In this method, initially Chebyshev nodes are suitably chosen in the desired ECG interval and then these nodes are utilized by Lagrange interpolation method for restoration of ECG signals. The results of the interpolation are further improved by segmenting the ECG signals into suitable number of segments using Bottom-Up time series segmentation approach. The performance of the proposed method is analyzed in terms of standard ECG performance parameters: mean absolute deviation, root mean square deviation, percentage root mean square difference error, signal to noise ratio and cross correlation. Results obtained are compared with existing methods and are found to be superior and diagnostically acceptable. Practical implementation areas of the proposed method are also explored.