An approach to QRS complex detection based on multiscale mathematical morphology

Abstract

Accurate detection of electrocardiograph feature points is always critical for medical decision-making. A newly developed multi-scale mathematical morphology and its use in QRS complex detection was studied in this paper. It was found that the existing multi-scale mathematical morphology has a problem of a significant shift in feature points positioning. Aiming for exact position detection of feature points, an improved approach based on the concept of multi-scale mathematical morphology is proposed in this study. Experimental results show a much smaller shift in point positioning, and thus demonstrate the effectiveness of the proposed approach.