Scholarlink

The purpose of this study was the detection of QRSs in ECG signals using the Matching Pursuit (MP) method. Five groups of patients were considered: healthy controls (NORM), patients with ST-T change (STTC), myocardial infarction (MI), conduction disturbance (CD), and hypertrophy (HYP). The ECG signals were decomposed using MP algorithm in set of functions, so called atoms, characterized by specific energy, amplitude, frequency and scale, corresponding to different ECG patterns, QRS complexes and T waves. A threshold for frequency ranging from 6 Hz to 45 Hz and scale value from 0.02 to 0.069 was applied to identify the QRS patterns. The proposed algorithm allowed for the identification of QRSs with a high performance independently on the patient group. The average values of sensitivity, specificity, and positive and negative predictive values were 98.94%, 100%, 100%, and 99.30%, respectively. The QRSs detection using an algorithm based on MP decomposition and on the application of two thresholds for frequency and scale is very efficient, and it can be applied in ECG wearable devices for monitoring patients in real time.

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

QRS Detection Based on the Matching Pursuit Algorithm

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Mohammed Owahidur Rahman,
  • Piotr Augustyniak,
  • Elzbieta Olejarczyk

摘要

The purpose of this study was the detection of QRSs in ECG signals using the Matching Pursuit (MP) method. Five groups of patients were considered: healthy controls (NORM), patients with ST-T change (STTC), myocardial infarction (MI), conduction disturbance (CD), and hypertrophy (HYP). The ECG signals were decomposed using MP algorithm in set of functions, so called atoms, characterized by specific energy, amplitude, frequency and scale, corresponding to different ECG patterns, QRS complexes and T waves. A threshold for frequency ranging from 6 Hz to 45 Hz and scale value from 0.02 to 0.069 was applied to identify the QRS patterns. The proposed algorithm allowed for the identification of QRSs with a high performance independently on the patient group. The average values of sensitivity, specificity, and positive and negative predictive values were 98.94%, 100%, 100%, and 99.30%, respectively. The QRSs detection using an algorithm based on MP decomposition and on the application of two thresholds for frequency and scale is very efficient, and it can be applied in ECG wearable devices for monitoring patients in real time.