Preprocessing and Denoising Techniques for Electrocardiography and Magnetocardiography: A Review

Area of Science:

• Biomedical Engineering
• Cardiology
• Signal Processing

Background:

• Electrocardiography (ECG) and Magnetocardiography (MCG) are vital for cardiovascular disease (CVD) detection.
• Both ECG and MCG signals are prone to various noise interferences, impacting diagnostic accuracy.
• A systematic comparison of ECG and MCG denoising techniques is lacking.

Purpose of the Study:

• To systematically review and analyze advancements in ECG and MCG signal preprocessing over the last decade.
• To categorize and compare ECG denoising methods based on noise types (baseline wander, EMG, PLI, composite noise).
• To examine MCG signal denoising challenges and explore the complementary nature of ECG and MCG.

Main Methods:

• Systematic literature review of preprocessing techniques for ECG and MCG signals.
• Categorization of denoising methods based on noise characteristics and signal type.
• Comparative analysis of traditional filtering, machine learning, and hybrid approaches.

Main Results:

• Identified various denoising techniques for ECG, addressing specific noise types.
• Highlighted the complexities and challenges in MCG signal denoising.
• Emphasized the complementary roles of ECG and MCG and the potential of MCG in enhancing CVD diagnosis.
• Evaluated limitations of current denoising methods in clinical settings.

Conclusions:

• Traditional filtering methods remain relevant but hybrid strategies combining machine learning offer significant potential.
• Future directions include explainable and multi-task neural networks for improved denoising and diagnostic accuracy.
• This review provides a framework for selecting and enhancing denoising techniques to improve CVD diagnostics.