Classification of Electrocardiogram Using Color Images with Pixel Method by Deep CNN

Abstract

This research mainly focuses on the electrocardiogram (ECG) signal classification using Deep Neural Network and continuous wavelet transformation (CWT) in MATLAB. It contains theoretical derivations and parameters equations with design and examples. First, we will explain types of ECG signal classification, ECG signal database used for classification, how we can convert ECG signal to image using CWT Scalogram, Transfer learning via pre-trained AlexNet deep convolutional neural network (CNN). After that, explain the MATLAB code for this database creation, which means conversion of 1D ECG signal to CWT scalogram image. Finally, explain the MATLAB code for training and validation of AlexNet for classification. All the necessary graphs are included, and this work's limitations and future scope are also discussed.