Key result
The SimEMG database was successfully created, comprising 147 single-lead ECG recordings (37 noise-free and 110 noise-contaminated) with an average SNR of 8.53 dB to enable direct evaluation of EMG noise removal methods.
Why the study?
Evaluating ECG noise-removal techniques has relied on ad hoc additive mixtures due to the lack of a standard database for testing and comparing denoising methods.
The SimEMG database provides the first open-source dataset of simultaneously recorded genuine EMG-noise-free and -contaminated ECG signals to evaluate denoising algorithms.
Answers are grounded in this paper and cite their sources.
Absence of standardized ECG-EMG noise databases limits reliable denoising validation; leaves open consensus benchmarks for clinical ECG use.
Goal:Noise on recorded electrocardiographic (ECG) signals may affect their clinical interpretation. Electromyographic (EMG) noise spectrally coincides with the QRS complex, which makes its removal particularly challenging. The problem of evaluating the noise-removal techniques has commonly been approached by algorithm testing on the contaminated ECG signals constructedad hocas an additive mixture of a noise-free ECG signal and noise. Consequently, there is an absence of a unique/standard database for testing and comparing different denoising methods. We present a SimEMG database recorded by a novel acquisition method that allows for direct recording of the genuine EMG-noise-free and -contaminated ECG signals. The database is available as open source.
No takes yet. Share an insight, caveat, or question.
Atanasoski et al. (2023) studied Healthy (n=14). SimEMG acquisition method was evaluated on Average signal-to-noise ratio (SNR) of noise-contaminated signals. The SimEMG database was successfully created, comprising 147 single-lead ECG recordings (37 noise-free and 110 noise-contaminated) with an average SNR of 8.53 dB to enable direct evaluation of EMG noise removal methods.
Synapse has enriched 10 related papers. Consider them for comparative context: