Heart Rate Detection on Single-Arm ECG by Using Dual-Median Approach

ORHAN U., AYDIN A.

ARABIAN JOURNAL FOR SCIENCE AND ENGINEERING, cilt.45, sa.8, ss.6573-6581, 2020 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 45 Sayı: 8
  • Basım Tarihi: 2020
  • Doi Numarası: 10.1007/s13369-020-04574-8
  • Dergi Adı: ARABIAN JOURNAL FOR SCIENCE AND ENGINEERING
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aerospace Database, Communication Abstracts, Metadex, Pollution Abstracts, zbMATH, Civil Engineering Abstracts
  • Sayfa Sayıları: ss.6573-6581
  • Anahtar Kelimeler: Single-arm ECG, Heart rate detection, Dual median, IIR filter, Wavelet transform, QRS DETECTION, RATE-VARIABILITY, CLASSIFICATION
  • Çukurova Üniversitesi Adresli: Evet

Özet

This study introduces a system to obtain correct heart rate values on single-arm ECG signals. For preparing a dataset, single-arm and dual-arm measurements were taken from the healthy volunteers; then, to increase the diversity of the data and test the limits of the method, some artificial noises are added to the unfiltered raw signals. The signals in the dataset were cleaned with the IIR filter, and then, their significant frequency sub-bands were selected using wavelet transform. The heart rate (HR) determined by the proposed algorithm from the single arm and compared with the ones of the dual-arm signals. The median function, which is used twice in the algorithm, did not only help to overcome the artifacts, but also facilitate to count peaks. According to the results, if the algorithm used in real-time applications, it is enough to wait only for 5 s from the moment that the system first started to find the HR with an MAE around one bpm, and 1 min for an MAE around 0.6 bpm. The most important advantage of this study is the superior performance of the algorithm on finding HR from noisy single-arm ECG signals with the lowest error in the literature and using dual-arm ECG measurement ground truth and as a reference to optimize the parameters of the algorithm.