[From the abstract of this article] Use of bandpass filter as a preprocessing for Hilbert transform in electrocardiographic QRS complex feature detection is discussed here, based on computer simulation re- sults. The main topic is onset and offset estimation on QRS com- plex. The bandpass filter is configured for the frequency range of QRS complex. Hilbert transform is applied to the filter output in order to generate an envelope signal which most reflects the QRS complex. From such an envelope signal, the onset and off- set of each QRS complex are estimated by means of the tangent line projection method. The simulation ECG waveforms were generated from a dipole source vector with the lead vectors of the standard limb leads as an application of the lead field the- ory and the single dipole source model. The simulation results demonstrated that the estimation was stable against changes in the ECG source-measurement relation, which depends on the ECG lead selection and the ECG lead vector.
