Scholarlink

<p>The QRS complex is an important component in ECG signals. Current QRS detection methods struggle with varying ECG morphologies and noise, leading to errors and the need for parameter adjustments. This paper introduces a novel, parameter-free QRS detection algorithm, which utilizes the cumulative contributions of different frequencies at any given time using continuous wavelet transform to remove noise without parameter adjustment. A moving average threshold is then applied for QRS detection. The performance of the proposed method was assessed on challenging ECG data and compared with popular existing methods, demonstrating high sensitivity (Se%) and positive predictivity (+ P%) across five diverse databases: MIT-BIH Arrhythmia Database (Se%: 99.84, +P%: 99.78), MIT-BIH Noise Stress Database (Se%: 94.74, +P%: 87.84), Glasgow University Database (Se%: 97.86%, +P%: 98.89%), China Physiological Signal Challenge 2020 Database (Se%: 99.28, +P%: 97.35), and China Physiological Signal Challenge 2019 Database (Se%: 95.92, +P%: 93.85). Notably, It outperformed existing algorithms in the most comprehensive and challenging dataset (China Physiological Signal Challenge 2019) by at least 6.46% in sensitivity and 2.08% in positive predictivity. The novelty of the proposed method lies in its adaptation to noise and morphological changes without parameter tuning, making it robust across different ECG conditions.</p>

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

Robust and parameter free QRS complex detection in ECG for heart disease diagnosis and monitoring

  • 期刊论文

  • 发表日期：

    2025年9月5日

查看全文
  • Abdullah Alhalabi,
  • Saleh Alzahrani,
  • Mahbubunnabi Tamal

摘要

The QRS complex is an important component in ECG signals. Current QRS detection methods struggle with varying ECG morphologies and noise, leading to errors and the need for parameter adjustments. This paper introduces a novel, parameter-free QRS detection algorithm, which utilizes the cumulative contributions of different frequencies at any given time using continuous wavelet transform to remove noise without parameter adjustment. A moving average threshold is then applied for QRS detection. The performance of the proposed method was assessed on challenging ECG data and compared with popular existing methods, demonstrating high sensitivity (Se%) and positive predictivity (+ P%) across five diverse databases: MIT-BIH Arrhythmia Database (Se%: 99.84, +P%: 99.78), MIT-BIH Noise Stress Database (Se%: 94.74, +P%: 87.84), Glasgow University Database (Se%: 97.86%, +P%: 98.89%), China Physiological Signal Challenge 2020 Database (Se%: 99.28, +P%: 97.35), and China Physiological Signal Challenge 2019 Database (Se%: 95.92, +P%: 93.85). Notably, It outperformed existing algorithms in the most comprehensive and challenging dataset (China Physiological Signal Challenge 2019) by at least 6.46% in sensitivity and 2.08% in positive predictivity. The novelty of the proposed method lies in its adaptation to noise and morphological changes without parameter tuning, making it robust across different ECG conditions.