Scholarlink

<p>Analyzing physiological measurements such as heart rate requires an exact peak in a noise-corrupted photoplethysmograph (PPG) signal. When evaluated in an laboratory conditions that do not perfectly simulate living conditions, traditional approaches exhibit proficiency with noise-free photoplethysmography (PPG) signals; however, their robustness significantly diminishes in free-living conditions due to various noise interferences. This paper presents a dual-decoder structure combining U-net and Long-Short-Term-Memory (LSTM). In addition, we suggest a preprocessing strategy for enhancing PPG noise-resiliency and robust peak detection by transforming the signal into the frequency domain. We collect a dataset of preadolescents of 207 cases using Polar Verity and Polar H10 during the MAST process. The proposed method shows a precision score of peak detection, 89%. In contrast, peak detection performance of the PPG signal is 27% because of false peaks and noise. We proves to be accurate for detecting PPG peaks even in the presence of noise. Furthermore, it can contribute to analyzing PPG signals corrupted by daily noise.</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

Deep Learning Based Method to Electrocardiogram Reconstruction via Photoplethysmography

  • 期刊论文

  • 发表日期：

    2025年6月30日

查看全文
  • Changhun Seo,
  • Hyeokhyeon Kwon,
  • Jiyeong Baek,
  • Sangrok Lee,
  • Jihye Lee,
  • Yeseul Lee,
  • Susanna Joo,
  • Hyoun K. Kim,
  • Changmin Lee

摘要

Analyzing physiological measurements such as heart rate requires an exact peak in a noise-corrupted photoplethysmograph (PPG) signal. When evaluated in an laboratory conditions that do not perfectly simulate living conditions, traditional approaches exhibit proficiency with noise-free photoplethysmography (PPG) signals; however, their robustness significantly diminishes in free-living conditions due to various noise interferences. This paper presents a dual-decoder structure combining U-net and Long-Short-Term-Memory (LSTM). In addition, we suggest a preprocessing strategy for enhancing PPG noise-resiliency and robust peak detection by transforming the signal into the frequency domain. We collect a dataset of preadolescents of 207 cases using Polar Verity and Polar H10 during the MAST process. The proposed method shows a precision score of peak detection, 89%. In contrast, peak detection performance of the PPG signal is 27% because of false peaks and noise. We proves to be accurate for detecting PPG peaks even in the presence of noise. Furthermore, it can contribute to analyzing PPG signals corrupted by daily noise.