ECG Signal Denoising and Features Extraction Using Unbiased FIR Smoothing.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Electrocardiography (ECG) signal feature extraction is crucial for diagnosing heart conditions.
• Artefacts and noise in ECG signals impede accurate feature extraction.
• Linear prediction is a standard but not always optimal technique for ECG signal processing.

Purpose of the Study:

• To develop an adaptive unbiased finite impulse response (UFIR) filter for improved ECG signal denoising and feature extraction.
• To enhance the detection of key ECG waveform parameters (P-wave, QRS-complex, T-wave).
• To address limitations of traditional linear prediction methods in ECG analysis.

Main Methods:

• Implementation of a p-shift unbiased finite impulse response (UFIR) filter with p < 0 for smoothing.
• Development of an adaptive averaging horizon for the UFIR filter, adjusting to signal dynamics.
• Evaluation of the algorithm's performance using real ECG data from normal heartbeats.

Main Results:

• The adaptive UFIR algorithm demonstrated superior denoising capabilities compared to standard linear predictors and UFIR filters.
• The developed filter achieved improved signal-noise ratio (SNR) for feature extraction.
• Accurate detection of P-wave, QRS-complex, and T-wave durations and amplitudes was achieved.

Conclusions:

• The adaptive UFIR filter offers a more efficient and effective approach to ECG signal processing than traditional methods.
• This algorithm improves the accuracy of extracting diagnostic features from noisy ECG signals.
• The method shows promise for enhanced detection of heart abnormalities through precise ECG waveform analysis.