Weighted averaging of ECG signal based on the number of turning points

In the case of real ECG processing, because of the nonstationary EMG noise, weighted averaging (WA) should be used instead of classical averaging (CA). Weights may be calculated from the noise variances of the isoelectronic sections of ECG. Instead of this badly defined parameter we propose an integral parameter of the whole heart cycle. The number of turning points of the hysteresis filtered signal proved to be appropriate for WA.