Scholarlink

Cardiac arrhythmia is a significant heart disorder distinguished by abnormal heart rhythms. It can cause severe health conditions if it is undetected. Electrocardiogram (ECG) signals provide vital information about heart function which makes them a valuable tool for arrhythmia detection. In this study, a novel strategy is proposed for the classification of cardiac arrhythmias using advanced deep learning techniques, CNN-LSTM-Attention model specifically a Convolutional Neural Network (CNN) combined with Long Short-Term Memory (LSTM) layers as well as an attention mechanism. The CNN model automatically extracts key spatial features from digitized ECG datasets, the LSTM model captures temporal dependencies and patterns in the sequential data which ensures a more comprehensive analysis of heart activity and attention mechanism further enhancing the model by assigning higher importance to relevant time steps in the ECG sequence. This model is evaluated on the CPEIC cardiac dataset public dataset. The proposed model achieves superior performance in terms of accuracy of 99.29%, precision of 99.29%, recall of 99.2%, and F1-score 97.87% compared to the existing traditional methods. This method is a hybrid approach that leverages the strengths of both CNN and Bi-LSTM, offering a robust solution for real-time and automated cardiac arrhythmia detection.

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

Enhancing Cardiac Arrhythmia Detection: Deep Learning-Based CNN-LSTM-Attention Framework for Digitized ECG Classification

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Israt Jahan,
  • Warda Ruhin Parsub,
  • Md. Samiul Islam,
  • Rifat Ara Rimi,
  • Fariha Mosharof,
  • Ahmed Wasif Reza

摘要

Cardiac arrhythmia is a significant heart disorder distinguished by abnormal heart rhythms. It can cause severe health conditions if it is undetected. Electrocardiogram (ECG) signals provide vital information about heart function which makes them a valuable tool for arrhythmia detection. In this study, a novel strategy is proposed for the classification of cardiac arrhythmias using advanced deep learning techniques, CNN-LSTM-Attention model specifically a Convolutional Neural Network (CNN) combined with Long Short-Term Memory (LSTM) layers as well as an attention mechanism. The CNN model automatically extracts key spatial features from digitized ECG datasets, the LSTM model captures temporal dependencies and patterns in the sequential data which ensures a more comprehensive analysis of heart activity and attention mechanism further enhancing the model by assigning higher importance to relevant time steps in the ECG sequence. This model is evaluated on the CPEIC cardiac dataset public dataset. The proposed model achieves superior performance in terms of accuracy of 99.29%, precision of 99.29%, recall of 99.2%, and F1-score 97.87% compared to the existing traditional methods. This method is a hybrid approach that leverages the strengths of both CNN and Bi-LSTM, offering a robust solution for real-time and automated cardiac arrhythmia detection.