Optimal Selection of Wavelet Function and Decomposition Level for Removal of ECG Signal Artifacts

摘要

Electrocardiogram (ECG) signals are used for prognosis of anatomic heart functions in clinical routine. During recording, ECG is contaminated with artifacts owing to specific reasons, which makes accurate delineation and interpretation of ECG more difficult. The commonly noticed artifacts include; baseline wanders, power line interference and muscle tremors. This paper aims at elucidation for the algorithms for optimal selection of wavelet function and decomposition level for removal of these artifacts in ECG, while preserving the diagnostic information. The features of wavelet transforms viz, multi-resolution, information in time and frequency, and thresholding have been employed for the implementation of algorithms in MATLAB. The performance of wavelet functions at various decomposition levels has been evaluated using percentage root mean square difference (PRO) criterion. Simulated results indicate that optimal value of wavelet function is db7 at 8th decomposition level for removal of artificially added baseline wander with PRO of 0.1636%, coif2 at 5th decomposition level for removal of artificially added power line interference with PRO of 0.3980% and db7 at 8th decomposition level for removal of artificially added muscle tremor with PRO of 0.8959% in selected records of European ST-T database (EDB). The proposed method shows better performance than existing methods in terms of lower PRO. Moreover it also addresses to the issues of conventional methods.