ECG-signal multi-classification model based on squeeze-and-excitation residual neural networks

Abstract

Background Accurate electrocardiogram (ECG) interpretation is crucial in the clinical ECG workflow because it is most likely associated with a disease that can cause major problems in the body. In this study, we developed an ECG-signal multi-classification model using deep learning. Methods We used a squeeze-and-excitation residual network (SE-ResNet), which is used for efficient image classification. Experiments were performed for seven different types of lead-II ECG data obtained from the Korea University Anam Hospital in South Korea. These seven types are normal sinus rhythm, atrial fibrillation, atrial flutter, sinus bradycardia, sinus tachycardia, premature ventricular contraction and first degree atrioventricular block. We compared SE-ResNet with a residual network (ResNet), as a baseline model. Results The SE-ResNet classifier with 152 layers achieved F1 scores of 97.05% for the seven-class classifications. Our model surpassed the baseline model, ResNet, by + 1.40% for the seven-class classifications. Conclusion For ECG-signal multi-classification, considering the F1 scores SE-ResNet might be better than the ResNet baseline model.

Convolutional Neural Network for Drowsiness Detection Using EEG Signals

“…Simple CNN, ResNet, WaveNet, and Inception are among the best CNNs networks widely used in biomedical signals analysis studies. Based on recent works [ 42 , 83 , 84 , 85 , 86 , 87 , 88 , 89 , 90 , 91 , 92 , 93 , 94 , 95 , 96 , 97 ], a comparative analysis is provided in the following using various performance criteria as complexity , 1D-dimension , performance and time-consumption . In this regard, specific three tests (2, 3 and 4 states) with various values are given for each criterion as following.…”

An ECG Signal Classification Method Based on Dilated Causal Convolution

“…To verify the superiority of the proposed method in running time, we reproduced the network model used in [ 10 – 12 , 22 – 24 ] and recorded the running time of the model on the testing set. Table 3 shows the running time of the four different models in the testing set.…”

“…[ 23 ] gets the highest classification accuracy in the comparative experiments. But due to the complex network structure, [ 23 ] takes the longest time in the testing set, with a duration of 48.37 s. And because of the deeper network depth, the running time of [ 24 ] on the testing set is 46.20s.…”

“…[ 23 ] uses time-frequency features to process the original data, and an artificial neural network (ANN) is used as feature extractors and classifiers. In literature [ 24 ], the authors use the squeeze-and-excitation residual network (SE-ResNet) to detect abnormal occurrence.…”

ECG Heartbeat Classification Based on an Improved ResNet-18 Model

“…For instance, Zhou et al [ 20 ] proposed an attention mechanism based on ResNet for ECG data processing, using two commonly used databases—the MIT-BIH database and the Physikalisch-Technische Bundesanstalt diagnostic ECG database (PTB), which stands out in both databases. Park et al [ 21 ] proposed a SE-ResNet, a residual network with a squeeze-and-excitation block, which outperforms the ResNet baseline model. Han et al [ 22 ] used a multilead residual neural network (ML-ResNet) with three residual blocks and feature fusion to detect and locate myocardial infarction using 12 leads ECG records.…”