Alexander fractional differential window filter for ECG denoising

Area of Science:

• Biomedical Engineering
• Signal Processing
• Applied Mathematics

Background:

• Electrocardiogram (ECG) signals are vital for monitoring heart activity but are susceptible to noise during recording and transmission.
• Noise in ECG signals can obscure important diagnostic information, hindering accurate anomaly detection.
• Effective ECG signal denoising is crucial for reliable cardiac diagnostics.

Purpose of the Study:

• To propose and evaluate a novel denoising filter, the Alexander fractional differential window (AFDW) filter, for ECG signals.
• To assess the performance of the AFDW filter in removing various types of noise from ECG data.
• To compare the AFDW filter's efficacy against existing state-of-the-art denoising algorithms.

Main Methods:

• The Alexander fractional differential window (AFDW) filter was designed using concepts from generalized Alexander polynomials and fractional calculus (R-L differential equation).
• A forward filter was formulated, and a backward filter was constructed by reversing its coefficients; the AFDW filter is the average of these.
• Performance was quantified using Signal-to-Noise Ratio (SNR), Mean Squared Error (MSE), Wavelet Energy based Diagnostic Distortion (WEDD), and a novel Morphological Power Preservation Measure (MPPM).

Main Results:

• The AFDW filter achieved high average SNR (e.g., 22.01 dB for power line noise) and low MSE across different noise types.
• The proposed Morphological Power Preservation Measure (MPPM) demonstrated that the AFDW filter preserves signal power and QRS morphology effectively.
• Extensive testing on the MIT-BIH arrhythmia database, including combined and real-world noisy ECG signals, showed superior performance compared to existing methods.

Conclusions:

• The Alexander fractional differential window (AFDW) filter is a highly effective tool for denoising electrocardiogram (ECG) signals.
• The AFDW filter demonstrates superior performance in noise reduction while preserving essential signal characteristics, including QRS morphology.
• The proposed AFDW filter represents a significant advancement in ECG signal processing and offers a robust solution for clinical applications.