Automated ECG Segmentation in Telehealth System: A Case Study with Real Brazilian Data

The segmentation of electrocardiogram (ECG) signals is essential for accurately analyzing critical features such as the P wave, the QRS complex, and the T wave, aiding in the diagnosis of heart disorders.This study explores Deep learning techniques applied to data from Telehealth Network of Minas Gerais (TNMG).The model achieved a precision of 0.88 (P wave), 0.91 (QRS), and 0.93 (T wave), with recall of 0.81 (P wave), 0.91 (QRS), and 0.85 (T wave), and F1-scores of 0.84 (P wave), 0.91 (QRS), and 0.84 (T wave), resulting in an overall accuracy of 91.14%.With the Hospital das Clnicas in Belo Horizonte processing thousands of ECG exams daily, automated segmentation can speed up triage and improve decision-making, contributing to faster and more efficient diagnoses within the Brazilian public health system.