AI-Enabled Atrial Fibrillation Detection, Prediction, and Classification Based on ECG Data: A Review
TL;DR
This review systematically examines AI-based approaches for AF analysis, evaluating their advantages and limitations using key performance metrics, including accuracy, sensitivity, specificity, and F-score.
Abstract
Atrial Fibrillation (AF) is a prevalent and serious cardiac arrhythmia associated with significant health risks, including stroke and heart failure. Traditional AF detection using electrocardiogram (ECG) recordings relies on manual interpretation, which is time-consuming and prone to human error. Advances in artificial intelligence (AI) have enabled automated AF detection, prediction, and classification, leveraging deep learning models such as Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks, alongside traditional machine learning classifiers like Random Forest (RF).This review systematically examines AI-based approaches for AF analysis, evaluating their advantages and limitations using key performance metrics, including accuracy, sensitivity, specificity, and F-score. Additionally, it discusses challenges such as data variability, sensor limitations, and the integration of AI into clinical workflows. The latest advancements in hybrid models and domain adaptation techniques are explored, addressing key constraints related to model generalizability and data quality. Future research directions emphasize the importance of explainable AI, continuous monitoring through wearable technologies, and clinical validation through real-world trials. While AI holds transformative potential for AF diagnosis and management, significant technical and clinical challenges remain for its widespread adoption in healthcare.
