1D Convolutional Neural Network Based ECG Classification System for Cardiovascular Disease Detection

EasyChair Preprint 6066

Abstract

The study on cardiovascular disease has always been a popular medical topic around the world. This paper presents a deep learning (DL) method based on a convolutional neural network (CNN) algorithm to identify patients' cardiovascular arrhythmia by using a multi-lead ECG signals. In addition to the input and output layers, the proposed CNN model includes six layers, i.e., two convolution layers, two pooling layers and two fully connected layers within a residual block. The focus of this work is to classify the ECG signals into five classes; namely, Left Bundle Branch Block (LBBB), Right Bundle Branch Block (RBBB), Atrial Premature Contraction (APC), Premature Ventricular Contraction (PVC) and Normal beat(N).  We evaluated the proposed method by using the MIT-BIH arrhythmia dataset. According to the results, our proposed method achieved an average accuracy of 97.8% for the classification of 13,200 instances.

Keyphrases: Cardiovascular disease, Convolutional Neural Network (CNN), Electrocardiogram (ECG), Healthcare, deep learning

@booklet{EasyChair:6066,
  author    = {Hassen Mahdhaoui and Jamel Hattay and Hela Haj Mohamed and Samir Belaid},
  title     = {1D Convolutional Neural Network Based ECG Classification System  for Cardiovascular Disease Detection},
  howpublished = {EasyChair Preprint 6066},
  year      = {EasyChair, 2021}}