Abstract
Electrocardiogram (ECG) contains the rhythmic features of continuous heartbeat and morphological features of ECG waveforms and varies among different diseases. Based on ECG signal features, we propose a combination of multiple neural networks, the multichannel parallel neural network (MLCNN-BiLSTM), to explore feature information contained in ECG. The MLCNN channel is used in extracting the morphological features of ECG waveforms. Compared with traditional convolutional neural network (CNN), the MLCNN can accurately extract strong relevant information on multilead ECG while ignoring irrelevant information. It is suitable for the special structures of multilead ECG. The Bidirectional Long Short-Term Memory (BiLSTM) channel is used in extracting the rhythmic features of ECG continuous heartbeat. Finally, by initializing the core threshold parameters and using the backpropagation algorithm to update automatically, the weighted fusion of the temporal-spatial features extracted from multiple channels in parallel is used in exploring the sensitivity of different cardiovascular diseases to morphological and rhythmic features. Experimental results show that the accuracy rate of multiple cardiovascular diseases is 87.81%, sensitivity is 86.00%, and specificity is 87.76%. We proposed the MLCNN-BiLSTM neural network that can be used as the first-round screening tool for clinical diagnosis of ECG.
Highlights
ECG signal is a periodic and nonstationary time series biological signal
To explore the rhythmic features contained in ECG and improve the accuracy of classification and recognition, we propose a combination of multiple neural networks and multichannel parallel network model (MLCNN-Bidirectional Long Short-Term Memory (BiLSTM)). e structure is shown in Figure 4. e preprocessed 12-lead ECG data with a duration of 10 seconds are inputted into the input layer of MLCNN channel as a two-dimensional array for the extraction of the morphological features of ECG waveforms
MLCNN is more suitable for multilead ECG data classification and recognition. e experimental results show that the MLCNN-BiLSTM proposed in this paper has the best performance among several models, its ACC is 87.81%, and it is effective in dealing with ECG classification
Summary
ECG signal is a periodic and nonstationary time series biological signal. It is based on the principle that cardiac electrical activity presents a complex nonlinear dynamic system [1]. Auxiliary diagnosis based on deep learning is data-oriented and can automatically extract ECG features through a model, mine complex association patterns and rich information from data, effectively prevent uncertainty in manual feature extraction, and provide a feasible technical way for ECG big data analysis. Differences among ECG are usually addressed by using deep learning algorithms in mining ECG morphological and rhythmic features [5]. In this study, based on the above viewpoints, we propose a combination of multiple neural networks, the multichannel parallel neural network model (MLCNN-BiLSTM), and use it in the multiclass classification disease recognition of multilead ECG. E sensitivity of different cardiovascular diseases to morphology and rhythm is automatically learned, and the multiclass classification auxiliary diagnosis model is established. E experimental results show that the proposed combination between multiple neural networks and multichannel parallel neural network (MLCNN-BiLSTM) achieves good results in multilead ECG classification. e effectiveness of disease screening is verified through visual analysis of the highdimensional features of the model after time-space fusion
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
