Suppression of electromyogram interference on the electrocardiogram by transform domain denoising

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Electromyogram (EMG) interference is a common artifact in electrocardiogram (ECG) recordings.
• This interference can obscure diagnostically important ECG waveform segments.
• Existing methods may not adequately address non-stationary Gaussian noise characteristics of EMG.

Purpose of the Study:

• To develop and validate a novel method for suppressing electromyogram (EMG) interference in electrocardiogram (ECG) recordings.
• To improve the signal-to-noise ratio (SNR) and reduce amplitude errors in ECG data.
• To preserve the integrity of ECG waveforms for accurate clinical interpretation.

Main Methods:

• Assumed EMG as long-term non-stationary Gaussian noise.
• Employed two successive decompositions and data transformation for Wiener filtering.
• Rearranged and aligned successive ECG cycles by R-wave, forming a matrix.
• Applied short-window discrete cosine transform (DCT) for inter-cycle de-correlation.
• Utilized translation-invariant wavelet domain Wiener filtering for intra-cycle de-correlation.

Main Results:

• Achieved an improvement in signal-to-noise ratio (SNR) exceeding 10 dB.
• Demonstrated a threefold reduction in mean relative amplitude errors.
• Successfully reduced ripple artifacts around signal transients.
• Preserved diagnostically important waveform segments in ECG signals.

Conclusions:

• The proposed method effectively suppresses EMG interference in ECG recordings.
• The technique enhances ECG signal quality, improving diagnostic accuracy.
• This approach offers a robust solution for artifact removal in clinical electrophysiology.