An ECG denoising technique based on AHIN block and gradient difference max loss.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Electrocardiogram (ECG) signals are prone to noise interference, leading to information loss.
• Existing deep learning models for ECG denoising lack robustness and may overlook signal gradient differences.

Purpose of the Study:

• To propose a novel deep learning denoising method for ECG signals.
• To enhance the robustness and accuracy of ECG signal processing.

Main Methods:

• Developed a two-stage denoising approach using an attention half instance normalization (AHIN) block.
• Introduced a gradient difference max loss (GDM Loss) function to minimize waveform distortion.
• Validated the method through extensive experiments comparing it with state-of-the-art techniques.

Main Results:

• The proposed method achieved excellent performance in signal-to-noise ratio (SNR) and root-mean-square error (RMSE).
• Demonstrated significant improvements in noise reduction under various noise types (BW, MA, EM).
• The method effectively minimizes information loss and corrects waveform distortions.

Conclusions:

• The novel deep learning approach offers a robust and effective solution for ECG signal denoising.
• The proposed AHIN block and GDM Loss function contribute to superior performance in preserving signal integrity.
• This advancement has the potential to improve the accuracy of ECG-based diagnostics.