Wavelet-Based Heart-Rate Detection And Ecg Classification Of Arrhythmia Using Alexnet Deep Cnn

Abstract

Effective monitoring of heart patients based on heart signals has the potential to save a large number of lives. Classification and prediction of heart diseases based on ECG signals have become increasingly important for patients and doctors over the last decade. The majority of contemporary techniques are based on custom-designed features for automatic heart rate detection and ECG signal classification. The fundamental objective of this thesis work is to calculation of heart-rate and create a comprehensive learning-based technique that can classify ECG signals more accurately. We collected all data from PhysioNet for doing this work. We calculated heart rate using the Discrete Wavelet Transform (DWT) and trained a pre-trained Convolutional Neural Network (CNN), specifically AlexNet, to classify ECG signals. To begin, we extracted a spectrogram for each of the signals and transformed them to RGB images using the Continuous Wavelet Transform (CWT); these images were then put into AlexNet and trained with minor specification adjustments. The study's findings indicate that our technique achieves an accuracy of 97.14 %