Scholarlink

This paper proposes a deep learning-based approach for arrhythmia classification in electrocardiogram (ECG) signals using a self-attention mechanism. The model leverages a combination of Convolutional Neural Networks (CNN) for feature extraction and Long Short-Term Memory (LSTM) networks to capture temporal dependencies in the ECG data. To enhance the model's ability to focus on relevant features, the scaled dot product, a self-attention mechanism is integrated within the LSTM framework, allowing the model to assign varying levels of importance to different time steps in the signal. Experimental results demonstrate that the incorporation of attention mechanisms significantly improves classification performance, with accuracy 98% in four categories of arrhythmias, offering an effective solution for the automated detection of arrhythmia in clinical settings.

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

Arrhythmia Classification of ECG Signal Using Deep Learning with Self-attention

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Cong Tung Dinh,
  • Cheong Siew Ann,
  • Van Huan Vu,
  • Le Minh Nguyen,
  • Ngoc Dung Bui

摘要

This paper proposes a deep learning-based approach for arrhythmia classification in electrocardiogram (ECG) signals using a self-attention mechanism. The model leverages a combination of Convolutional Neural Networks (CNN) for feature extraction and Long Short-Term Memory (LSTM) networks to capture temporal dependencies in the ECG data. To enhance the model's ability to focus on relevant features, the scaled dot product, a self-attention mechanism is integrated within the LSTM framework, allowing the model to assign varying levels of importance to different time steps in the signal. Experimental results demonstrate that the incorporation of attention mechanisms significantly improves classification performance, with accuracy 98% in four categories of arrhythmias, offering an effective solution for the automated detection of arrhythmia in clinical settings.