Scale selection technique for heartbeat detection algorithm

This paper presents an empirical study that aims to determine the frequency band best suited for ECG heartbeat detection operation for portable monitoring electronic device. This study was accomplished using the Continuous Wavelet Transform technique and in particular the Mexican-hat wavelet. This study was performed to minimize the effect of motion artifacts type of noise as it is prevalent and difficult to remove. It consists of gradually tuning the scale of the wavelet and observing its effect on the output signal and more precisely its effect on both real beats and peaks that correspond to motion artifacts noise. The result concluded was that the best wavelet to be used for such a task was this having a frequency response centered in the value 32 Hz.