Scholarlink

<p>A computationally efficient information technology for the automatic detection of <i>R</i>-peaks in electrocardiogram signals is proposed, designed for use in portable long-term monitoring systems. A piecewise polynomial approximation method based on second-order Hermite polynomials is applied to transform a discrete sequence of samples into a continuous analytical representation of the signal. This ensures the continuity of both the approximation function and its first derivative, enabling their analytical computation. This approach minimizes noise impact and ensures accurate <i>R</i>-peak localization despite significant baseline drift and motion artifacts. The mathematical framework is integrated into a comprehensive clinical decision support system using a microservice architecture with vector-based visualization. The results of testing on the MIT-BIH Arrhythmia Database open dataset showed the method’s high efficiency, with an average F-score exceeding 99.1%, and in some complex cases, 100% accuracy was achieved. The technology’s low computational cost makes it suitable for real-time cardiac monitoring in embedded systems and Internet of Things (IoT) devices.</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

Accurate R-Peak Detection in Electrocardiograms: Piecewise Polynomial Approximation, Noise Reduction, and First-Derivative Analysis

  • 期刊论文

  • 发表日期：

    2026年8月1日

查看全文
  • M. S. Yefremov,
  • A. V. Liashko,
  • O. B. Stelia,
  • B. V. Batsak,
  • Iu. V. Krak

摘要

A computationally efficient information technology for the automatic detection of R-peaks in electrocardiogram signals is proposed, designed for use in portable long-term monitoring systems. A piecewise polynomial approximation method based on second-order Hermite polynomials is applied to transform a discrete sequence of samples into a continuous analytical representation of the signal. This ensures the continuity of both the approximation function and its first derivative, enabling their analytical computation. This approach minimizes noise impact and ensures accurate R-peak localization despite significant baseline drift and motion artifacts. The mathematical framework is integrated into a comprehensive clinical decision support system using a microservice architecture with vector-based visualization. The results of testing on the MIT-BIH Arrhythmia Database open dataset showed the method’s high efficiency, with an average F-score exceeding 99.1%, and in some complex cases, 100% accuracy was achieved. The technology’s low computational cost makes it suitable for real-time cardiac monitoring in embedded systems and Internet of Things (IoT) devices.