Opportunities and challenges of deep learning methods for electrocardiogram data: A systematic review.

https://research-information.bris.ac.uk/ws/files/225112566/Manuscript_CBM.pdf

Application of deep learning techniques for heartbeats detection using ECG signals-analysis and review

Background:The electrocardiogram (ECG) is one of the most commonly used diagnostic tools in medicine and healthcare. Deep learning methods have achieved promising results on predictive healthcare tasks using ECG signals. Objective:This paper presents a systematic review of deep learning methods for ECG data from both modeling and application perspectives. Methods:We extracted papers that applied deep learning (deep neural network) models to ECG data that were published between Jan. 1st of 2010 and Feb. 29th of 2020 from Google Scholar, PubMed, and the DBLP. We then analyzed each article according to three factors: tasks, models, and data. Finally, we discuss open challenges and unsolved problems in this area. Results: The total number of papers extracted was 191. Among these papers, 108 were published after 2019. Different deep learning architectures have been used in various ECG analytics tasks, such as disease detection/classification, annotation/localization, sleep staging, biometric human identification, and denoising. Conclusion: The number of works on deep learning for ECG data has grown explosively in recent years. Such works have achieved accuracy comparable to that of traditional feature-based approaches and ensembles of multiple approaches can achieve even better results. Specifically, we found that a hybrid architecture of a convolutional neural network and recurrent neural network ensemble using expert features yields the best results. However, there are some new challenges and problems related to interpretability, scalability, and efficiency that must be addressed. Furthermore, it is also worth investigating new applications from the perspectives of datasets and methods. Significance: This paper summarizes existing deep learning research using ECG data from multiple perspectives and highlights existing challenges and problems to identify potential future research directions.

Opportunities and Challenges of Deep Learning Methods for Electrocardiogram Data: A Systematic Review

Cardiac arrhythmia is associated with abnormal electrical activities of the heart, which can be reflected by altered characteristics of electrocardiogram (ECG). Due to the simplicity and non-invasive nature, the ECG has been widely used for detecting arrhythmias and there is an urgent need for automatic ECG detection. Up to date, some algorithms have been proposed for automatic classification of cardiac arrhythmias based on the features of the ECG; however, their stratification rate is still poor due to unreliable features of signal characteristics or limited generalization capability of the classifier, and therefore, it remains a challenge for automatic diagnosis of arrhythmias. In this paper, we propose a new method for automatic classification of arrhythmias based on deep neural networks (DNNs). The two DNN models constitutive of residual convolutional modules and bidirectional long short-term memory (LSTM) layers are trained to extract features from raw ECG signals. The extracted features are concatenated to form a feature vector which is trained to do the final classification. The algorithm is evaluated based on the test set of China Physiological Signal Challenge (CPSC) dataset with F1 measure regarded as the harmonic mean between the precision and recall. The resulting overall F1 score is 0.806, FAF score is 0.914 for atrial fibrillation (AF), FBlock score is 0.879 for block, FPC and FST scores are 0.801 and 0.742 for premature contraction and ST-segment change, which demonstrates a good performance that may have potential practical applications.

/pdf/automatic-cardiac-arrhythmia-classification-using-31nce8t50b.pdf

Automatic Cardiac Arrhythmia Classification Using Combination of Deep Residual Network and Bidirectional LSTM

To reduce the high mortality rate from cardiovascular disease (CVD), the electrocardiogram (ECG) beat plays a significant role in computer-aided arrhythmia diagnosis systems. However, the complex variations and imbalance of ECG beats make this a challenging issue. Since ECG beat data exist in heavily imbalanced category, an effective long short-term memory (LSTM) recurrence network model with focal loss (FL) is proposed. For this purpose, the LSTM network can disentangle the timing features in complex ECG signals, while the FL is used to resolve the category imbalance by downweighting easily identified normal ECG examples. The advantages of the proposed network have been verified in the MIT-BIH arrhythmia database. Experimental results show that the LSTM network with FL achieved a reliable solution to the problem of imbalanced datasets in ECG beat classification and was not sensitive to quality of ECG signals. The proposed method can be deployed in telemedicine scenarios to assist cardiologists into more accurately and objectively diagnosing ECG signals.

/pdf/an-effective-lstm-recurrent-network-to-detect-arrhythmia-on-4nu1k2urlr.pdf

An Effective LSTM Recurrent Network to Detect Arrhythmia on Imbalanced ECG Dataset

This paper introduces a novel deep learning-based algorithm that integrates a long short-term memory (LSTM)-based auto-encoder (AE) network with support vector machine (SVM) for electrocardiogram (ECG) arrhythmias classification. The LSTM-based AE network (LSTM-AE) is used to learn the features from ECG arrhythmias signals, and the SVM is used to classify those signals from the learned features. The LSTM-AE consists of an encoder model, which extracts high-level feature information from ECG arrhythmias signals through LSTM network, and a decoder model which outputs reconstruct ECG arrhythmias signals from high-level features through LSTM network. Experiments show that the proposed method can learn better features than the traditional method without any prior knowledge, presenting a good potential for the ECG arrhythmias classification. In the classification of five heartbeats types, including normal, left bundle branch block (LBBB), right bundle branch block (RBBB), atrial premature complexes (APC), premature ventricular contractions (PVC), the proposed method achieved average accuracy, sensitivity, and specificity of 99.74%, 99.35%, and 99.84%, respectively, in the beat-based cross-validation approach, and 85.20%, 62.99%, and 90.75%, respectively, in the record-based cross-validation approach, in public MIT-BIH Arrhythmia Database. While based on the Advancement of Medical Instrumentation (AAMI) standards, the proposed method achieved average accuracy, sensitivity, and specificity of 99.45%, 98.63%, and 99.66%, respectively, in the beat-based cross-validation approach.

LSTM-Based Auto-Encoder Model for ECG Arrhythmias Classification

This paper presents a novel deep learning-based method that integrates a Convolutional Auto-Encoder (CAE) with support vector machine (SVM) for finger-vein verification. The CAE is used to learn the features from finger-vein images, and the SVM is used to classify finger vein from these learned feature codes. The CAE consists of a finger-vein encoder, which extracts high-level feature representation from raw pixels of the images, and a decoder which outputs reconstruct finger-vein images from high-level feature code. As an effective classifier, SVM is introduced in this paper to classify the feature code which is obtained from CAE. Experiments prove that the proposed deep learning-based approach has superior performance in learning features than traditional method without any prior knowledge, presenting a good potential in the verification of finger vein.

Convolutional Autoencoder Model for Finger-Vein Verification

This paper aims at proposing an abnormality detection framework for electrocardiogram (ECG) signals, which owns unbalance distribution among different classes and gaining high accuracy in rhythm/morphology abnormalities classification. The proposed framework is composed of two models: data augmentation model and classification model. In this framework, data augmentation model is designed to recast a class-balanced training dataset by generating artificial data of minor class. The outputs of augmentation model are transferred into classification model. The classification model is designed to identify abnormalities accurately after training using both the experimental and generated datasets. Data augmentation model is supported by auxiliary classifier generative adversarial network (ACGAN). We construct Generator and Discriminator of the ACGAN by stacking multiple 1-dimensional convolutional layers with small kernel size. Dropout function and batch normalization are added to prevent gradients vanish and speed up convergence. In order to evaluate the performance of augmentation model, a set of quantitative indicators are introduced to verify the quality of generated ECG signals. We establish classification model based on stacked residual network parallel connected with long short-term memory (LSTM) network. The experimental study is conducted for single heartbeat detection and consecutive heartbeat detection. The results based on standard benchmark, MIT-BIH, and competition database provided by 2018 China physiological signal challenge (CPSC) have verified the proposed framework can achieve high performance in robustness and accuracy for class-imbalanced dataset.

/pdf/ecg-arrhythmias-detection-using-auxiliary-classifier-1pcqhwi2zk.pdf

ECG Arrhythmias Detection Using Auxiliary Classifier Generative Adversarial Network and Residual Network

This article proposes a new loss function termed arccosine center loss, which can learn interclass and intraclass information simultaneously, to improve the discriminative ability of convolutional neural networks for finger vein verification. Specifically, the purpose of arccosine center loss is to reduce intraclass distance and increase the interclass distance through network model training. With the combination of softmax loss and arccosine center loss, the proposed network model can extract features with the interclass dispensation and intraclass compactness, which improves the discriminative ability of learned features. Experimental studies have confirmed the effectiveness and efficiency of the proposed method for finger vein verification.

ArcVein-Arccosine Center Loss for Finger Vein Verification

This paper presents a novel deep learning-based method that integrates a Convolutional Auto-Encoder (CAE) with Convolutional Neural Network (CNN) for finger vein verification. The CAE is used to learn the feature codes from finger vein images and the CNN is used to classify finger vein from these learned feature codes. The CAE consists of a finger vein encoder, which extracts high-level feature representation from raw pixels of the images, and a decoder which outputs reconstruct finger vein images from high-level feature code. Experimental study proves that the proposed deep learning based method has superior performance in learning features than traditional method without any prior knowledge, presenting a good potential in the verification of finger vein.

Borui Hou

Papers

LSTM-Based Auto-Encoder Model for ECG Arrhythmias Classification

Convolutional Autoencoder Model for Finger-Vein Verification

ECG Arrhythmias Detection Using Auxiliary Classifier Generative Adversarial Network and Residual Network

ArcVein-Arccosine Center Loss for Finger Vein Verification

Convolutional Auto-Encoder Based Deep Feature Learning for Finger-Vein Verification