Abstract:
This project is based on Machine Learning on an ECG signal in real time. The ECG
signal is passed through some defined procedures that allow removal of noise. This
After data is then passed through a thresholding algorithm used to detect R peaks,
that Q and S peaks found, respectively. These features are than passed to a
classifier that classifies these signals accordingly. The ECG signal is taken from a
are
pulse sensor connected to a Raspberry pi through an ADC. The procedures were first
applied on an ECG dataset provided by Physionet. The project shows results of real
time classification that can be further improved. The classifier shows an accuracy of
75%
