An sEMG Denoising Method with Improved Threshold Estimation for Rapid Keystroke Tasks

TL;DR

This paper introduces a new sEMG denoising method that improves signal quality during rapid keystrokes by optimizing decomposition and thresholding techniques.

Contribution

The novel contribution is an improved threshold estimation method integrated with WO-optimized VMD for sEMG denoising in high-duty-cycle tasks.

Findings

The proposed method achieved a median SNR improvement (ΔSNR) of 2.75 to 6.65 dB across various input SNRs.

It reduced root-mean-square error (RMSE) by 27% to 53% while maintaining spectral fidelity with a median frequency variation rate below 3.48%.

The method outperformed wavelet, EMD, EMD-IT, and FCN baselines in denoising rapid keystroke sEMG signals.

Abstract

What are the main findings?
Developed a WO-VMD-based denoising framework for rapid keystroke sEMG, where WO adaptively optimizes key VMD parameters to enhance decomposition stability.Proposed an improved threshold estimation for rapid keystroke sEMG to prevent over-suppression induced by thresholding.

Developed a WO-VMD-based denoising framework for rapid keystroke sEMG, where WO adaptively optimizes key VMD parameters to enhance decomposition stability.

Proposed an improved threshold estimation for rapid keystroke sEMG to prevent over-suppression induced by thresholding.

What are the implications of the main findings?
Validated on 18 subjects (0–15 dB input SNR), achieving higher ΔSNR and ΔRMSE% than wavelet, EMD, EMD-IT and FCN baselines.Provided an “adaptive decomposition + physiologically informed thresholding” approach applicable to rapid keystroke sEMG analysis and other nonstationary biomedical signals.

Validated on 18 subjects (0–15 dB input SNR), achieving higher ΔSNR and ΔRMSE% than wavelet, EMD, EMD-IT and FCN baselines.

Provided an “adaptive decomposition + physiologically informed thresholding” approach applicable to rapid keystroke sEMG analysis and other nonstationary biomedical signals.

Surface electromyography (sEMG) signals are inevitably affected by noise during acquisition, thereby degrading signal quality and analytical reliability. Most existing denoising methods combine signal decomposition with thresholding, and their performance depends on empirically set decomposition parameters and threshold estimation. However, in high-rate repetitive motions such as rapid keystrokes, sustained high-duty-cycle muscle activation biases universal-threshold noise estimation, leading to unreliable thresholds. To overcome these issues, an sEMG denoising method that integrates the Walrus Optimizer (WO) with Variational Mode Decomposition (VMD) is proposed. WO is employed to optimize key VMD parameters, including the number of modes K and the penalty factor α. Based on this method, an improved threshold estimation strategy is developed to accommodate high-duty-cycle sEMG during rapid keystrokes. It reduces thresholding-induced over-attenuation of meaningful myoelectric components. The dataset included 18 participants with sEMG recorded from six muscles during rapid keystroke tasks (10 trials per participant; 20 keystrokes per trial). Across input signal-to-noise ratios (SNRs) of 0, 5, 10, 15 dB, the proposed method achieved a median SNR improvement (ΔSNR) ranging from 2.75 to 6.65 dB and a median root-mean-square error (RMSE) reduction rate (ΔRMSE%) ranging from 27% to 53%, while maintaining spectral fidelity with a median of median frequency variation rate (ΔMDF%) below 3.48%.These results indicate that the proposed method provides an efficient and reliable solution for sEMG signal processing in rapid keystroke analysis.

Figures