Scholarlink

Cardiovascular diseases are among the most serious conditions to diagnose in the healthcare system. Arrhythmias is a type of cardiovascular disorder that involve irregularities in the heart’s rhythm and lead to severe complications if not detected. The purpose of this study is to compare two deep learning models Convolutional Neural Network (CNN) versus Bidirectional Long-Short Term Memory (Bi-LSTM) for the categorization of cardiac rhythm disorders. The proposed models were trained, validated, and evaluated using Electrocardiogram (ECG) recordings from the MIT-BIH database, with heartbeats annotated into multiple categories, including Nodal Escape Beat, Paced Beat, Right Bundle Branch Block Beat, Left Bundle Branch Block Beat, Ventricular Premature Contraction, Supraventricular Premature Beat, Normal Beat, and others. To analyze the performance of the models, the confusion matrix used to compute accuracy, precision, recall and F1-score. Experimental results demonstrate slightly high performance for the CNN with overall accuracy of 99.71%, and 99.37% for the Bi-LSTM.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Comparative Study: CNN Versus Bi-LSTM Models for Cardiac Arrhythmias Classification

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Oumaima Gamgami,
  • Reda Korikache,
  • Amine Chaieb

摘要

Cardiovascular diseases are among the most serious conditions to diagnose in the healthcare system. Arrhythmias is a type of cardiovascular disorder that involve irregularities in the heart’s rhythm and lead to severe complications if not detected. The purpose of this study is to compare two deep learning models Convolutional Neural Network (CNN) versus Bidirectional Long-Short Term Memory (Bi-LSTM) for the categorization of cardiac rhythm disorders. The proposed models were trained, validated, and evaluated using Electrocardiogram (ECG) recordings from the MIT-BIH database, with heartbeats annotated into multiple categories, including Nodal Escape Beat, Paced Beat, Right Bundle Branch Block Beat, Left Bundle Branch Block Beat, Ventricular Premature Contraction, Supraventricular Premature Beat, Normal Beat, and others. To analyze the performance of the models, the confusion matrix used to compute accuracy, precision, recall and F1-score. Experimental results demonstrate slightly high performance for the CNN with overall accuracy of 99.71%, and 99.37% for the Bi-LSTM.