A review of arrhythmia detection based on electrocardiogram with artificial intelligence

Area of Science:

• Cardiology
• Biomedical Engineering
• Artificial Intelligence

Background:

• Portable electrocardiogram (ECG) devices are increasing ECG diagnoses.
• Traditional computer-aided arrhythmia diagnosis faces limitations in data quality and expert knowledge.
• Artificial intelligence (AI) offers potential for high-precision arrhythmia diagnosis and early screening.

Purpose of the Study:

• To describe AI applications in arrhythmia detection.
• To summarize the advantages and limitations of various AI approaches.
• To provide guidance for future AI research in arrhythmia detection.

Main Methods:

• Machine learning (ML) and deep learning (DL) for ECG signal processing.
• DL for automatic feature extraction and classification of ECG signals.
• AI integration into smart ECG devices for screening.

Main Results:

• AI methods effectively address ECG signal denoising, quality assessment, and wave delineation.
• DL algorithms automatically learn features for accurate heartbeat and rhythm classification.
• AI-powered systems can significantly reduce physician workload in ECG analysis.

Conclusions:

• AI, particularly ML and DL, shows significant promise in advancing arrhythmia detection.
• AI integration into devices facilitates widespread early arrhythmia screening.
• Further research into AI methods will refine diagnostic accuracy and efficiency.