An efficient algorithm to remove electrocardiogram from diaphragmatic electromyography by using wavelet spectrum and secant method

An efficient algorithm based on wavelet spectrum and secant method is proposed, investigated and exploited for removing electrocardiogram(ECG) from diaphragmatic electromyography(EMGdi). As a strong interference, ECG has the high amplitude and spectrum overlap in raw EMGdi. Inspired from the characteristic of ECG, low frequency wavelet spectrum of raw EMGdi and gate function are used for fixing the peak position of QRS wave. After calculating the undisturbed average energy at the neighborhood of the R peak in the field of wavelet spectrum, the threshold adaptive criterion employs undisturbed energy to match the energy of ECG interference interval and shrink those interference coefficients. The optimal thresholds in each ECG interference are derived from secant method. In comparing with other de-noising algorithms, the processed results of simulated EMGdi signal and clinical respiratory signals show that the wavelet spectrum and secant method proposed in this paper can remove ECG interference in respiratory signals effectively and reserve its signal characteristics satisfyingly.