Baseline wander removal in ECG and AHA recommendations

Baseline wander is a kind of noise that affects all ECG signals. We have recently proposed a novel approach to its removal which is based on Quadratic Variation Reduction (QVR). The approach has very favorable properties and has shown to be effective in removing baseline wander, while preserving the ST segment level. It requires the determination of a detrending parameter. In this paper, we derive a linear time-invariant filter approximating QVR. The filter retains (approximately) the same optimality properties as QVR. Moreover, it provides a criterion for choosing the proper value of the parameter governing QVR, as a function of the spectral characteristics of baseline wander noise. Simulation results show that the filter is effective in removing baseline wander, while introducing minor distortion in the ST segment.