Processing and Feature Extraction of ECG Signals for Arrhythmia Detection in AI Models

Abstract

Arrhythmias, such as tachycardia and bradycardia, are prevalent in postoperative patients, especially within the first week after surgery. These conditions can lead to significant health risks, particularly in settings with inadequate monitoring resources, such as rural areas in Peru. This paper proposes an advanced approach to arrhythmia detection using ECG signal processing and feature extraction to feed artificial intelligence (AI) models. By optimizing training data and comparing ECG signal features, this method aims to enhance the accuracy and efficiency of arrhythmia identification, thereby improving patient monitoring and reducing associated health risks. Postoperative patients are at high risk of developing heart pathologies, which generally manifest as persistent arrhythmias, especially after the first week after surgery; conditions that might lead to significant health risks. Modern heart monitoring systems should prevent patients and their family when arrhythmias are present, thus they attend to medical assistance. This paper proposes an advanced approach to arrhythmia detection using ECG signal processing and feature extraction to feed artificial intelligence models. This method aims to enhance the accuracy and efficiency of arrhythmia detection, and thereby improving patients’ monitoring systems, by optimizing training data.

Tasks

Reproductions