ECG Classification Using Artificial Neural Networks

Abstract

We propose two Artificial Neural Networks (ANN) architectures for classification of electrocardiogram (ECG) signals to compare effectiveness between them. The atrial fibrillation (AF) classification data set provided by PhysioNet/CinC Challenge 2017 was used. The ANNs proposed are a feed forward neural network (FFNN) and a convolutional neural network (CNN). In order to train the convolutional neural network we transformed the ECG signals to images. The convolutional neural network overcomes the other by reaching an average accuracy of 97.6% in prediction set.