Wavelet based optimized polynomial threshold function for ECG signal denoising

The Electrocardiogram (ECG) signal is a bio-signal which provides shows the electrical activity of the heart and provides information about the heart´s condition. Data retrieval from ECG signal becomes a tough task when the signal is corrupted with noise. In this paper, denoising of ECG signal in wavelet domain using a polynomial based threshold function is proposed. The coefficients of the function are optimized using artificial bee colony (ABC) algorithm. The proposed function is compared with three existing functions, hard, soft and non-negative garrote threshold function. The performance parameters used are output signal to noise ratio, mean square error and cross correlation coefficient. Performance is compared for two different ECG signals.