Scholarlink

Cardiovascular diseases (CVDs) represent a leading factor in global mortality rates, posing substantial challenges for healthcare systems worldwide, presenting significant healthcare challenges in Metropolitan Lima. In Peru, delays and limited technology within the healthcare system hinder the early detection and timely treatment of CVDs, contributing to elevated mortality rates. This paper presents an “Intelligent Monitoring and Alert System” designed to address these gaps through a mobile alert application and an integrated web platform powered by deep learning and IoT-driven data collection. The system leverages a ResNet model to classify ECG data in real time, with Fog Computing ensuring rapid processing of metrics from sensors, such as the AD8232. Initial results demonstrate robust performance, achieving an AUC of 0.84 for ECG classification from a single-lead. The system has undergone validation according to GAMP 5 standards, along with usability testing conducted with healthcare professionals and received favorable usability ratings (4.53/5), with results demonstrating the system’s effectiveness and utility.

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

Intelligent Monitoring and Analysis System for the Early Detection of Cardiovascular Anomalies Based on IoT and Deep Learning

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Dominik Mendoza Ramos,
  • Cristina Vidal Falcon,
  • Abel Rosales Caururu

摘要

Cardiovascular diseases (CVDs) represent a leading factor in global mortality rates, posing substantial challenges for healthcare systems worldwide, presenting significant healthcare challenges in Metropolitan Lima. In Peru, delays and limited technology within the healthcare system hinder the early detection and timely treatment of CVDs, contributing to elevated mortality rates. This paper presents an “Intelligent Monitoring and Alert System” designed to address these gaps through a mobile alert application and an integrated web platform powered by deep learning and IoT-driven data collection. The system leverages a ResNet model to classify ECG data in real time, with Fog Computing ensuring rapid processing of metrics from sensors, such as the AD8232. Initial results demonstrate robust performance, achieving an AUC of 0.84 for ECG classification from a single-lead. The system has undergone validation according to GAMP 5 standards, along with usability testing conducted with healthcare professionals and received favorable usability ratings (4.53/5), with results demonstrating the system’s effectiveness and utility.