Abstract
This study aims to explore a new deep learning strategy for electrocardiogram (ECG) denoising under adverse conditions of non-stationary power-line interference (PLI) with amplitude changes or nominal frequency deviations. The study presents an exhaustive training strategy of deep convolutional autoencoder (CAE), while input with one of the largest PhysioNet 12-lead ECG databases contaminated by simulated sinusoidal PLI noise with augmented settings. Twelve ECG leads (I, II, III, aVR, aVL, aVF, V1–V6) from 14890 PTB-XL records, divided patient-wise to training (50%, 7441 records), validation (20%, 2979 records) and test (30%, 4470 records) are superimposed by PLI with five signal-to-noise ratios (SNR) (–2.5, 0, 2.5, 5, 7.5 dB), nine frequencies (48, 48.5, 49, 49.5, 50, 50.5, 51, 51.5, 52 Hz), 12 amplitude slew rates ±(50, 100, 250, 500, 750, 1000 μV/s), and 14 frequency slew rates ±(0.01, 0.025, 0.05, 0.075, 0.1, 0.15, 0.2 Hz/s). CAE receptive field inputs one ECG lead with 1024 samples (2.048 s for 500 Hz sampling rate). CAE architecture is designed with seven 1D-convolutional layers, including three encoder layers (filters x kernel size = 16 × 8, 8 × 8, 8 × 8) and four decoder layers (8 × 8, 8 × 8, 16 × 8, 1 × 8) with linear activation function and same padding. CAE non-linear operations for max-pooling and up-sampling (pool size of 2) follow each encoder and decoder convolutional layers, respectively. Adam optimizer and mean squared error loss function are applied for CAE training over 250 epochs. The quality of clean ECG reconstruction in CAE output is evaluated by root-mean-square error (RMSE), percentage-root-mean-square difference (PRD) and improvement in signal-to-noise ratio (SNRimp). Statistical test results for denoising of all 12 ECG leads present median RMSE = 5.3 μV, PRD = 3.5%, SNRimp = 22–32 dB for SNR = –2.5 to 7.5 dB. The results do not substantially change for PLI frequencies 48–52 Hz, amplitude slew rates up to ±1000 μV/s and frequency slew rates up to ±0.2 Hz/s with median value divergence of \(\Delta \)RMSE < 2 \(\upmu{\text{V}} \), \(\Delta \)PRD < 1.5%, \(\Delta \)SNRimp ≤ 3 dB. The observed performance stability justifies the deep learning strategy for training a CAE with generalizable application for denoising of ECG signals with non-stationary PLI.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Bradley, W., et al.: Electrocardiogram in Clinical Medicine, 1st edn. John Wiley & Sons Ltd., Hoboken, NJ, US (2020). https://doi.org/10.1002/9781118754511
Sörnmo, L., Laguna P.: Chapter 6 – the electrocardiogram—a brief background. Bioelectr. Signal Process. Card. Neurol. Appl. 2005, 411–452 (2005)
Dobrev, D., Neycheva, T.: High-quality biopotential acquisition without a reference electrode: power-line interference reduction by adaptive impedance balancing in a mixed analog–digital design. Med. Biol. Eng. Comput. 60, 1801–1814 (2022)
IEC 60601-2-25:2011: Medical Electrical Equipment - Part 2-25: particular requirements for the basic safety and essential performance of electrocardiographs, 2nd edn., International Electrotechnical Commission, pp. 1–196 (2011)
Dobrev, D., Neycheva, T., Mudrov, N.: High-Q comb filter for mains interference suppression. Annu. J. Electron. 3(1), 47–49 (2009)
Tabakov, S., Iliev, I., Krasteva, V.: Online digital filter and QRS detector applicable in low resource ECG monitoring systems. Ann. Biomed. Eng. 36(11), 1805–1815 (2008)
Levkov, C., Mihov, G., Ivanov, R., et al.: Removal of power-line interference from the ECG: a review of the subtraction procedure. Biomed. Eng. OnLine 4, 50 (2005). https://doi.org/10.1186/1475-925X-4-50
Dotsinsky, I., Stoyanov, T., Mihov, G.: Power-line interference removal from high sampled ECG signals using modified version of the subtraction procedure. Int. J. Bioautom. 24(4), 381–392 (2020)
Stoyanov, T., Christov, I., Jekova, I., Krasteva, V.: Online adaptive filter for mains interference suppression in diagnostic electrocardiographs: cases of amplitude and frequency deviation. Annu. J. Electron. 4, 150–153 (2010)
Rahman, M.Z.U., Karthik, G.V.S., Fathima, S.Y., Lay-Ekuakille, A.: An efficient cardiac signal enhancement using time–frequency realization of leaky adaptive noise cancelers for remote health monitoring systems. Measurement 46(10), 3815–3835 (2013)
Faiz, M., Kale, I.: Removal of multiple artifacts from ECG signal using cascaded multistage adaptive noise cancellers. Array 14, 100133 (2022)
Chiang, H., Hsieh, Y., Fu, S., et al.: Noise reduction in ECG signals using fully convolutional denoising autoencoders. IEEE Access 7, 60806–60813 (2019)
Dasan, E., Panneerselvam, I.: A novel dimensionality reduction approach for ECG signal via convolutional denoising autoencoder with LSTM. Biomed. Signal Process. Control 63, 102225 (2021)
Xiong, P., Wang, H., Liu, M., et al.: A stacked contractive denoising auto-encoder for ECG signal denoising. Physiol. Meas. 37, 2214–2230 (2016)
Casas, L., Klimmek, A., Navab, N., Belagiannis, V.: Adversarial signal denoising with encoder-decoder networks. In: 28th European Signal Processing Conference, pp. 1467–1471. IEEE, Amsterdam, Netherlands (2021)
Wagner, P., Strodthoff, N., Bousseljot, R.D., et al.: A large publicly available ECG dataset. Sci. Data 7, 154 (2020)
Němcová, A., Smíšek, R., Maršánová, L., Smital, L., Vítek, M.: A comparative analysis of methods for evaluation of ECG signal quality after compression. Biomed. Res. Int. 2018, 1868519 (2018)
Acknowledgements
This work was supported by the Bulgarian National Science Fund, grant КП-06-H42/3.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Ivanov, K., Jekova, I., Krasteva, V. (2023). Convolutional Autoencoder for Filtering of Power-Line Interference with Variable Amplitude and Frequency: Study of 12-Lead PTB-XL ECG Database. In: Sotirov, S., Pencheva, T., Kacprzyk, J., Atanassov, K.T., Sotirova, E., Ribagin, S. (eds) Recent Contributions to Bioinformatics and Biomedical Sciences and Engineering. BioInfoMed 2022. Lecture Notes in Networks and Systems, vol 658. Springer, Cham. https://doi.org/10.1007/978-3-031-31069-0_13
Download citation
DOI: https://doi.org/10.1007/978-3-031-31069-0_13
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-31068-3
Online ISBN: 978-3-031-31069-0
eBook Packages: EngineeringEngineering (R0)