Work

Design and use of a Denoising Convolutional Autoencoder for reconstructing electrocardiogram signals at super resolution

Year: 2024

Type: article

Abstract: Electrocardiogram (ECG) signals play a pivotal role in cardiovascular diagnostics, providing essential information on the heart's electrical activity. However, inherent noise and limited resolution can hinder accurate interpretation of ECG recordings. We present herein DCAE-SR a novel Denoising Convolutional AutoEncoder (DCAE) able to process ECG signals and produce a Super Resolution (SR) version. DCAE-SR receives an ECG window (of 5 seconds) sampled at 50 Hz (low resolution) as input and reconstructs a denoised super-resolution signal at 500 Hz. We applied the proposed DCAE-SR to publicly available PTB-XL dataset proving that DCAE-SR performs well in reconstructing high-resolution ECG signals from very low-resolution signals sampled at 50 Hz. Results have been compared with current state-of-the art methods for ECG super-resolution showing the efficacy of the proposed method. By using DCAE-SR it is possible to get a signal-to-noise ratio of 12.20 dB, mean square error of 0.0044, and root mean squared error of 4.86\%, a considerable better result than current state-of-the-art methods. We believe that DCAE-SR can be used to efficiently enhance information hidden in ECG signals, gathering heart related diseases.

Source: Artificial Intelligence in Medicine

DOI: 10.1016/j.artmed.2024.103058

Authors: Ugo Lomoio, Pierangelo Veltri, Pietro Hiram Guzzi, Pietro Liò

Institutions: Magna Graecia University, University of Calabria, University of Cambridge

Language: English

Topic: ECG Monitoring and Analysis

Subfield: Cardiology and Cardiovascular Medicine

Field: Medicine

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: hybrid

Estimated APC paid: $3,230