Wavelet Approach for ECG Baseline Wander Correction and Noise Reduction.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Electrocardiogram (ECG) analysis is crucial for drug safety and clinical diagnosis.
• ECG pre-processing requires correction of low-frequency baseline wander (BW) and high-frequency artifact noise.
• Existing methods may not adequately address both noise types simultaneously.

Purpose of the Study:

• To develop and present effective approaches for ECG baseline wander correction and de-noising.
• To utilize discrete wavelet transformation (DWT) for robust ECG signal pre-processing.
• To provide guidance on selecting DWT parameters for optimal noise reduction.

Main Methods:

• Baseline wander (BW) estimation using coarse approximation in DWT.
• Recommendations for wavelet selection and decomposition level optimization.
• High-frequency noise reduction via Empirical Bayes posterior median wavelet shrinkage with adaptive thresholding.

Main Results:

• The proposed DWT-based method effectively removes low-frequency baseline wander.
• High-frequency artifact noise is significantly reduced using wavelet shrinkage.
• Experimental application to a real ECG signal demonstrated successful noise elimination.

Conclusions:

• The presented DWT approaches provide an effective solution for ECG signal pre-processing.
• The methods enhance ECG data quality for improved diagnostic accuracy and drug development safety.
• This work offers practical guidelines for implementing DWT in ECG analysis.