Evolutionary optimization-based descendent adaptive filter for noise confiscation in electrocardiogram signals.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Computational Intelligence

Background:

• Electrocardiogram (ECG) signals are crucial for cardiac diagnostics but often suffer from noise and artefacts during recording.
• Signal degradation compromises the accuracy of heart condition assessment, necessitating effective noise cancellation techniques.

Purpose of the Study:

• To propose and evaluate swarm intelligence-based optimally tuned adaptive noise cancellers (ANCs) for denoising ECG signals.
• To compare the performance of ANCs optimized by the Seagull Optimization Algorithm (SOA), NLSHADE, and HGSA for artefact cancellation.

Main Methods:

• Development of ANCs optimized using SOA, NLSHADE, and HGSA for ECG signal denoising.
• Validation using public ECG datasets (ADB, QTDB) corrupted with additive white Gaussian noise at various SNRs.
• Performance evaluation using metrics like Percentage Root Mean Squared Deviation (PRD), Mean Squared Error (MSE), and Signal-to-Noise Ratio (SNR) improvement.

Main Results:

• The proposed SOA-assisted ANC achieved superior performance with a PRD of 3.40E-03 and MSE of 1.35E-11.
• An average SNR improvement of 10.986 dB was recorded, outperforming benchmark algorithms.
• Statistical validation using the Wilcoxon signed-rank test confirmed the effectiveness of the SOA-assisted ANC.

Conclusions:

• Swarm intelligence optimization, particularly SOA, offers a robust approach for enhancing ECG signal quality.
• The proposed ANCs effectively remove artefacts, improving diagnostic accuracy for cardiac conditions.
• This technique presents a significant advancement in biomedical signal processing for clinical applications.