A Cascaded Convolutional Neural Network for Assessing Signal Quality of Dynamic ECG

Abstract

Motion artifacts and myoelectrical noise are common issues complicating the collection and processing of dynamic electrocardiogram (ECG) signals. Recent signal quality studies have utilized a binary classification metric in which ECG samples are determined to either be clean or noisy. However, the clinical use of dynamic ECGs requires specific noise level classification for varying applications. Conventional signal processing methods, including waveform discrimination, are limited in their ability to remove motion artifacts and myoelectrical noise from dynamic ECGs. As such, a novel cascaded convolutional neural network (CNN) is proposed and demonstrated for application to the five-classification problem (low interference, mild motion artifacts, mild myoelectrical noise, severe motion artifacts, and severe myoelectrical noise). Specifically, this study finally categorizes dynamic ECG signals into three levels (low, mild, and severe) using the proposed CNN to meet clinical requirements. The network includes two components, the first of which was used to distinguish signal interference types, while the second was used to distinguish signal interference levels. This model does not require feature engineering, includes powerful nonlinear mapping capabilities, and is robust to varying noise types. Experimental data are composed of private dataset and public dataset, which were acquired from 90,000 four-second dynamic ECG signals and MIT-BIH Arrhythmia database, respectively. Experimental results produced an overall recognition rate of 92.7% on private dataset and 91.8% on public dataset. These results suggest the proposed technique to be a valuable new tool for dynamic ECG auxiliary diagnosis.

Artefact Detection in Impedance Pneumography Signals: A Machine Learning Approach

“…This decreases the need for prior knowledge and human effort in feature design. CNNs have been successfully used to detect artefacts in ECG, EEG and other (biomedical) signals [26][27][28].…”

A Deep Learning Approach for Featureless Robust Quality Assessment of Intermittent Atrial Fibrillation Recordings from Portable and Wearable Devices

“…However, this good performance drastically fell to 50% when AF recordings were considered in the study. A similar structure of two CNNs working in parallel was also proposed by Zhang et al [ 33 ] to identify three levels of noise (low, mild, and severe). The first network was one-dimensional and received original ECG as input, whereas the second one was two-dimensional and then fed with the ECG wavelet scalogram.…”

“…On the other hand, although most previous ECG quality indices have reported promising results on recordings from healthy subjects, their performance on signals acquired from patients with different pathological cardiac conditions has shown to be fairly limited [ 31 ]. This is the case of many algorithms whose ability to discern between clean and noisy ECG excerpts has been significantly decreased when dealing with ECG recordings obtained from patients with atrial arrhythmias, including AF [ 23 , 30 , 32 , 33 ]. Hence, the present work aims at introducing a novel algorithm for quality assessment of single-lead ECG recordings acquired with portable and wearable devices from patients with intermittent AF.…”

Comparative Study of Convolutional Neural Networks for ECG Quality Assessment

“…These networks are able to obtain the most relevant ECG characteristics without the need of detecting and delineating its fiducial points and other waveforms. To date, however, only a few works have introduced algorithms based on CNNs for ECG quality as-sessment [6,7]. Moreover, these methods could be notably overfitted, since they have been designed and trained from scratch with a reduced number of ECG samples.…”