ECG Classification Research Articles

Performance improvement of deep learning based multi-class ECG classification model using limited medical dataset

Performance improvement of deep learning based multi-class ECG classification model using limited medical dataset

Multi-lead ECG heartbeat classification of heart disease based on HOG local feature descriptor

IntroductionECG data play an important role in the diagnostics of various cardiovascular diseases. Classification of multi-lead ECG signals could be challenging even for well-trained physicians. In this study we propose a new approach for multi-lead ECG classification. MethodFive-types of 15-lead ECG data namely healthy control, bundle branch block, cardiomyopathy, Dysrhythmia, and myocardial infarction patients from two types of datasets, 5319 and 6647 heartbeats from Baqiyatallah and PTB Diagnostic ECG database, were used, respectively. One-dimensional total variation regularization was used to denoising ECG data. Heartbeats were extracted by one cardiologist and saved as images with jpg format. Histogram of oriented gradients method was used to extract feature of images. for classification task support vector machine and fully connected neural network were used. Five-fold cross validation was used for validating the models. ResultFor 15-lead ECG PTB Diagnostic database, the best classification models are SVM model with cubic (accuracy: 99.9%, Range: 99.77% - 100%) and quadratic (accuracy: 99.88%, Range: 99.77%-100%) kernel function, for this dataset fully connected accuracy is 99.4% with range of 99.02%- 99.70%. Regarding to the Baqyatallah dataset SVM with cubic (accuracy: 99.83%, Range:99.72%-100%) and quadratic (accuracy: 99.77%, Range: 99.62%-99.9%) were the best classification model and the accuracy for fully connected neural network was 99.1% with the range of 98.59%-99.62% based on HOG descriptors. Expected sigmodal kernel all classification method have accuracy more than 99%. Discussionsimultaneous use of HOG feature extraction method and appropriate classification algorithm such as SVM or fully connected neural network can classify 15-lead ECG heart-beat for different heart disease with high accuracy and adding other relevant patients’ information can be easily done in order to increase the method performance.

Non-stationary wavelet for ECG signal classification

Wavelet analysis has shown to be an interesting tool for representing ECG signals for classification. In this paper, we present a new ECG signal representation based on the notion of non-stationary wavelets. The main difference with the construction of standard wavelets is that the multiresolution spaces are generated by scale-dependent functions in order to achieve increased flexibility and sparseness. In order to customize the non-stationary wavelet to the given ECG classification task, we resort to the fireworks optimization algorithm, thus making the proposed method general and not constrained by the choice of a particular classifier. The proposed method is validated on AAMI classes of the well-known MIT data set. Results compared to standard stationary wavelets show a significant boost in accuracy.

Retracted: ECG‐ViT: A Transformer‐Based ECG Classifier for Energy‐Constraint Wearable Devices

Retracted: ECG‐ViT: A Transformer‐Based ECG Classifier for Energy‐Constraint Wearable Devices

Convolutional Neural Network-Based ECG-Assisted Diagnosis for Coal Workers.

To process and extract electrocardiogram (ECG, ECG, or EKG) features using a convolutional neural network (CNN) to establish an ECG-assisted diagnosis model. Coal workers who underwent physical examinations at Gequan Mine Hospital and Dongpang Mine Hospital of Hebei Jizhong Energy from July 2020 to September 2020 were selected as the study subjects. The ECG images were preprocessed. We use Python software and convolutional neural network to establish ECG images recognition and classification model.We usecalibration curve, calibration-in-the-large, Brier score, specificity, sensitivity, F1 score, Kappa value, accuracy, and area under the curve (AUC) of ROC to evaluate the performance of the model. The number of abnormal ECG results was 849, and the rate of abnormal results was 25.02%. The test set accuracies of the sinus bradycardia model, nonspecific intraventricular conduction delay model, myocardial ischemia model, and sinus tachycardia model were 97.66%, 96.49%, 93.62%, and 93.02%, respectively; sensitivities were 96.63%, 96.30%, 96.88% and 95.24%, respectively; specificities were 98.78%, 96.67%, 86.67%, and 90.90%, respectively; Brier scores were 0.03, 0.07, 0.09, and 0.11, respectively; Calibration-in-the-large values were 0.026, 0.110, 0.041, and 0.098, respectively. The convolutional neural network model can accurately identify the main ECG abnormality types of coal workers. Additionally, the main ECG abnormalities in these coal company workers were sinus bradycardia, non-specific intraventricular conduction delay, myocardial ischemia, and sinus tachycardia.

Improving ECG Classification Performance by Using an Optimized One-Dimensional Residual Network Model

Cardiovascular disease and its consequences on human health have never stopped and even show a trend of appearing in increasingly younger generations. The establishment of an excellent deep learning algorithm model to assist physicians in identifying and the early screening of ECG abnormalities can effectively improve the accuracy of diagnosis. Therefore, in this study, the deep residual network model is adapted for feature extraction and classification of ECG signals by pooling embedded into layers and double channel connection. At the same time, the wavelet adaptive threshold denoising algorithm is used to complete the high signal-to-noise filtering of ECG signals. Then, the alternate pooling residual network (APRN) is compared with the convolutional neural network (CNN), CNN with one residual unit (CNN-R), and the deep residual network (ResNet-18) using ECG datasets from the American MIT-BIH arrhythmia and ST segment abnormality database, European ST-T database, and sudden cardiac death ambulatory ECG database. The results are as follows: The average classification accuracy of the APRN on the four datasets is 97.89%, while the accuracies on CNN, CNN-R, and ResNet-18 are 97.17%, 97.53%, and 97.73%, respectively. In addition, compared with ResNet-18, the classification accuracy of our APRN on each class of data improves by 16.44% in total.

The ΔWaveECG: The differences to the normal 12‑lead ECG amplitudes

BackgroundThe QRS, ST segment, and T-wave waveforms of electrocardiogram are difficult to interpret, especially for non-ECG experts readers, like general practitioners. As the ECG waveforms are influenced by many factors, like body build, age, sex, electrode placement, even for experience ECG readers the waveform is difficult to interpret. In this research we have created a novel method to distinguish normal from abnormal ECG waveforms for an individual ECG based on the ECG amplitude distribution derived from normal standard 12‑lead ECG recordings. AimCreation of a normal ECG amplitude distribution to enable the distinction by non-ECG experts of normal from abnormal waveforms of the standard 12‑lead ECG. MethodsThe ECGs of healthy normal controls in the PTB-XL database were used to construct a normal amplitude distribution of the 12 lead ECG for males and females. All ECGs were resampled to have the same number of samples to enable the classification of an individual ECG as either normal or abnormal, i.e. within the normal amplitude distribution or outside, the ΔWaveECG. ResultsFrom the same PTB-XL database six ECG's were selected, normal, left and right bundle branch block, and three with a myocardial infarction. The normal ECG was obviously within the normal distribution, and all other five showed clear abnormal ECG amplitudes outside the normal distribution in any of the ECG segments (QRS, ST segment and remaining STT segment). ConclusionThe ΔWaveECG can distinguish the abnormal from normal ECG waveform segments, making the ECG easier to classify as normal or abnormal. Conduction disorders and ST changes due to ischemia and abnormal T-waves are effortless to detect, also by non-ECG expert readers, thus improving the early detection of cardiac patients.

Pragmatic analysis of ECG classification models & architectures from a statistical perspective

Electrocardiograms, also known as ECGs, are capable of representing a wide range of heart conditions, such as atrial fibrillation, arrhythmia, myopathy, and heart failure. Operations such as denoising, signal enhancement, feature extraction, feature selection, stratification, and post processing are developed in order to process ECG signals. Depending on the application that is being deployed, these processes may only work in their individual components or in combination with one another. Researchers over the years have developed a wide variety of algorithms, each specifically tailored to accomplish one of these tasks. The computational complexity of these algorithms, the number of diseases they can classify, the degree of accuracy they provide, the cost of deployment, and other factors can vary. For example, models that are based on convolutional neural networks (CNNs) have a high level of accuracy; however, these models are typically slow and highly complex to deploy, and as a result, theyare not used in clinical applications that have a low cost. However, linear classification models such as support vector machines (SVMs) have acceptable accuracy, a moderate level of complexity, but are lacking in terms of precision; as a result, they are not suitable for use in applications that require real-time processing. Because there is such a wide range of algorithms that are available, it is unclear to researchers which models would be the best fit for their application, which leads to an increase in both the cost of deployment and the amount of time needed to bring the product to market. The text that lies beneath this one examines several of the recently proposed ECG classification models and assesses them based on a number of statistical parameters, including computational delay, complexity of deployment, classification accuracy, precision, number of heart diseases covered, and so on.This is done in an effort to clarify the situation and remove some of the ambiguity. After completing this analysis, the text then compares their performance and provides an estimate of the models that are the most appropriate. Researchers now have the ability to select and implement the algorithmic models that are best suited for their deployments based on this review. In addition to that, this reviwe suggests a number of different enhancements that can be made to the models that have been examined, and researchers can implement these suggestions in order to increase the effectiveness of the models.

My Journey Through the Field of Medical Informatics.

In this contribution some achievements and milestones in the field of medical informatics, especially concerning decision support, as perceived by the author, are presented. The author focuses on those topics with respect to decision support that during his career in medical informatics impressed him and triggered him to convince his PhD students to start research on related topics. Both some of these achievements and the related research of some of his PhD students will be presented. The contribution starts with signal classification. Both ECG classification and sleep EEG classification are discussed. Then the use of Bayes' theorem for diagnostic purposes is discussed and some early applications pass review, among which the AAPHelp system developed by de Dombal and colleagues. Attention is subsequently paid to the advent of expert systems and other knowledge-based systems such as MYCIN and INTERNIST and to guideline-based decision support systems. Finally, the author presents his ideas about challenges for the field.

Arrhythmia Detection Based on WGAN-GP and SE-ResNet1D

A WGAN-GP-based ECG signal expansion and an SE-ResNet1D-based ECG classification method are proposed to address the problem of poor modeling results due to the imbalanced sample distribution of ECG data sets. The network architectures of WGAN-GP and SE-ResNet1D are designed according to the characteristics of ECG signals so that they can be better applied to the generation and classification of ECG signals. First, ECG data were generated using WGAN-GP on the MIT-BIH arrhythmia database to balance the dataset. Then, the experiments were performed using the AAMI category and inter-patient data partitioning principles, and classification experiments were performed using SE-ResNet1D on the imbalanced and balanced datasets, respectively, and compared with three networks, VGGNet, DenseNet and CNN+Bi-LSTM. The experimental results show that using WGAN-GP to balance the dataset can improve the accuracy and robustness of the model classification, and the proposed SE-ResNet1D outperforms the comparison model, with a precision of 95.80%, recall of 96.75% and an F1 measure of 96.27% on the balanced dataset. Our methods have the potential to be a useful diagnostic tool to assist cardiologists in the diagnosis of arrhythmias.

CBIR-DSS: Business Decision Oriented Content-Based Recommendation Model for E-Commerce

Business-based decision support systems have been proposed for a few decades in the e-commerce and textile industries. However, these Decision Support Systems (DSS) have not been so productive in terms of business decision delivery. In our proposed model, we introduce a content-based image retrieval model based on a DSS and recommendations system for the textile industry, either offline or online. We used the Fashion MNIST dataset developed by Zalando to train our deep learning model. Our proposed hybrid model can demonstrate how a DSS can be integrated with a system that can separate customers based on their personal characteristics in order to tailor recommendations of products using behavioral analytics, which is trained based on MBTI personality data and Deap EEG data containing numerous pre-trained EEG brain waves. With this hybrid, a DSS can also show product usage analytics. Our proposed model has achieved the maximum accuracy compared to other proposed state-of-the-art models due to its qualitative analysis. In the first section of our analysis, we used a deep learning algorithm to train our CBIR model based on different classifiers such as VGG-net, Inception-Net, and U-net which have achieved an accuracy of 98.2% with a 2% of minimized error rate. The result was validated using different performance metrics such as F-score, F-weight, Precision, and Recall. The second part of our model has been tested on different machine learning algorithms with an accuracy rate of 89.9%. Thus, the entire model has been trained, validated, and tested separately to gain maximum efficiency. Our proposal for a DSS system, which integrates several subsystems with distinct functional sets and several model subsystems, is what makes this study special. Customer preference is one of the major problems facing merchants in the textile industry. Additionally, it can be extremely difficult for retailers to predict customer interests and preferences to create products that fulfill those needs. The three innovations presented in this work are a conceptual model for personality characterization, utilizing an amalgamation of an ECG classification model, a suggestion for a textile image retrieval model using Denoising Auto-Encoder, and a language model based on the MBTI for customer rating. Additionally, we have proposed a section showing how blockchain integration in data pre-processing can enhance its security and AI-based software quality assurance in a multi-model system.

An Ultra-Energy-Efficient and High Accuracy ECG Classification Processor With SNN Inference Assisted by On-Chip ANN Learning.

The ECG classification processor is a key component in wearable intelligent ECG monitoring devices which monitor the ECG signals in real time and detect the abnormality automatically. The state-of-the-art ECG classification processors for wearable intelligent ECG monitoring devices are faced with two challenges, including ultra-low energy consumption demand and high classification accuracy demand against patient-to-patient variability. To address the above two challenges, in this work, an ultra-energy-efficient ECG classification processor with high classification accuracy is proposed. Several design techniques have been proposed, including a reconfigurable SNN/ANN inference architecture for reducing energy consumption while maintaining classification accuracy, a reconfigurable on-chip learning architecture for improving the classification accuracy against patent-to-patient variability, and a dual-purpose binary encoding scheme of ECG heartbeats for further reducing the energy consumption. Fabricated with a 28nm CMOS technology, the proposed design consumes extremely low classification energy (0.3μJ) while achieving high classification accuracy (97.36%) against patient-to-patient variability, outperforming several state-of-the-art designs.

Ensemble classification combining ResNet and handcrafted features with three-steps training

Objective. This work presents an ECG classifier for variable leads as a contribution to the Computing in Cardiology Challenge/CinC Challenge 2021. It aims to integrate deep and classic machine learning features into a single model, exploring the proper structure and training procedure. Approach. From the initial 88 253 signals, only 84 210 were included. Low quality and unscored recordings were excluded. Three different database subsets of 40 365 recording each were created by dividing in three normal sinus rhythm and sinus bradycardia recordings. Each subset was used to train a different model with shared architecture integrated as an ensemble to provide the final classification through major voting. Models contained a deep branch composed of a modified ResNet with dilation convolutional layers and squeeze and excitation Block that took as input windowed ECG signals. This was concatenated with a wide branch that integrated 20 cardiac rhythm features into a fully connected 3-layered network. Three different training steps were studied: just the deep branch (D), wide integration and training (D+W), and a final fine tuning of the deep branch posterior to wide training (D+W+D). Main Results. Results obtained in a local test set formed by a stratified 12.5% split of the given full dataset were presented for 2-lead and 12-lead models. The best training method was the 3-step D + W + D procedure obtaining a challenge metric of 0.709 and 0.677 for 12 and 2-lead models respectively. Significance. Integration of handcrafted features and deep learning model not only may increase the generalization capacity of the network but also provide a path to add explicit information into the classification decision process. To the best of our knowledge this is the first work studying the training procedure to properly integrate both types of information for ECG signals classification.

Issues in the automated classification of multilead ecgs using heterogeneous labels and populations

Objective. The standard twelve-lead electrocardiogram (ECG) is a widely used tool for monitoring cardiac function and diagnosing cardiac disorders. The development of smaller, lower-cost, and easier-to-use ECG devices may improve access to cardiac care in lower-resource environments, but the diagnostic potential of these devices is unclear. This work explores these issues through a public competition: the 2021 PhysioNet Challenge. In addition, we explore the potential for performance boosting through a meta-learning approach. Approach. We sourced 131,149 twelve-lead ECG recordings from ten international sources. We posted 88,253 annotated recordings as public training data and withheld the remaining recordings as hidden validation and test data. We challenged teams to submit containerized, open-source algorithms for diagnosing cardiac abnormalities using various ECG lead combinations, including the code for training their algorithms. We designed and scored the algorithms using an evaluation metric that captures the risks of different misdiagnoses for 30 conditions. After the Challenge, we implemented a semi-consensus voting model on all working algorithms. Main results. A total of 68 teams submitted 1,056 algorithms during the Challenge, providing a variety of automated approaches from both academia and industry. The performance differences across the different lead combinations were smaller than the performance differences across the different test databases, showing that generalizability posed a larger challenge to the algorithms than the choice of ECG leads. A voting model improved performance by 3.5%. Significance. The use of different ECG lead combinations allowed us to assess the diagnostic potential of reduced-lead ECG recordings, and the use of different data sources allowed us to assess the generalizability of the algorithms to diverse institutions and populations. The submission of working, open-source code for both training and testing and the use of a novel evaluation metric improved the reproducibility, generalizability, and applicability of the research conducted during the Challenge.

Inter-patient ECG classification with i-vector based unsupervised patient adaptation

This paper proposes an unsupervised patient adaptation approach to creating patient-specific deep neural network (DNN) classifiers for inter-patient ECG classification. The method exploits the information embedded in the patient-specific i-vectors derived from some unlabeled patient-specific ECG. The adaptation process comprises two stages of backpropagation (BP) fine-tuning, using the i-vector of a target patient as an auxiliary input to a middle layer of the DNN. In the first stage, labeled ECG data from a general population are used for creating a patient-adaptive DNN. Then, in the second stage, unlabeled ECG data from the target patient are used for further BP fine-tuning, using the labels hypothesized by the patient-adaptive DNN as the desired outputs. To ensure that only reliable data are used for adaptation, an information-theoretic heartbeat selector is employed to select the patients’ ECG with high-confidence hypothesized labels. Evaluations on the MIT-BIH arrhythmia dataset show that the proposed unsupervised adaptation leads to patient-specific ECG classifiers that outperform existing patient-specific models. The classifiers also perform comparably to patient-specific models obtained via supervised adaption. This unsupervised adaptation approach can fully automate patient adaptation, making personalized ECG classification more practical.

Lightweight Multireceptive Field CNN for 12-Lead ECG Signal Classification.

The electrical activity produced during the heartbeat is measured and recorded by an ECG. Cardiologists can interpret the ECG machine's signals and determine the heart's health condition and related causes of ECG signal abnormalities. However, cardiologist shortage is a challenge in both developing and developed countries. Moreover, the experience of a cardiologist matters in the accurate interpretation of the ECG signal, as the interpretation of ECG is quite tricky even for experienced doctors. Therefore, developing computer-aided ECG interpretation is required for its wide-reaching effect. 12-lead ECG generates a 1D signal with 12 channels among the well-known time-series data. Classical machine learning can develop automatic detection, but deep learning is more effective in the classification task. 1D-CNN is being widely used for CVDS detection from ECG datasets. However, adopting a deep learning model designed for computer vision can be problematic because of its massive parameters and the need for many samples to train. In many detection tasks ranging from semantic segmentation of medical images to time-series data classification, multireceptive field CNN has improved performance. Notably, the nature of the ECG dataset made performance improvement possible by using a multireceptive field CNN (MRF-CNN). Using MRF-CNN, it is possible to design a model that considers semantic context information within ECG signals with different sizes. As a result, this study has designed a multireceptive field CNN architecture for ECG classification. The proposed multireceptive field CNN architecture can improve the performance of ECG signal classification. We have achieved a 0.72 F1 score and 0.93 AUC for 5 superclasses, a 0.46 F1 score and 0.92 AUC for 20 subclasses, and a 0.31 F1 score and 0.92 AUC for all the diagnostic classes of the PTB-XL dataset.

A CONVOLUTIONAL NEURAL NETWORK FOR AUTOMATIC DISCRIMINATION OF PAUSE EPISODES DETECTED BY AN INSERTABLE CARDIAC MONITOR

A CONVOLUTIONAL NEURAL NETWORK FOR AUTOMATIC DISCRIMINATION OF PAUSE EPISODES DETECTED BY AN INSERTABLE CARDIAC MONITOR

Deep Learning-based ECG Classification on Raspberry PI using a Tensorflow Lite Model based on PTB-XL Dataset

The number of IoT devices in healthcare is expected to rise sharply due to increased demand since the COVID-19 pandemic. Deep learning and IoT devices are being employed to monitor body vitals and automate anomaly detection in clinical and non-clinical settings. Most of the current technology requires the transmission of raw data to a remote server, which is not efficient for resource-constrained IoT devices and embedded systems. Additionally, it is challenging to develop a machine learning model for ECG classification due to the lack of an extensive open public database. To an extent, to overcome this challenge PTB-XL dataset has been used. In this work, we have developed machine learning models to be deployed on Raspberry Pi. We present an evaluation of our TensorFlow Model with two classification classes. We also present the evaluation of the corresponding TensorFlow Lite FlatBuffers to demonstrate their minimal run-time requirements while maintaining acceptable accuracy.

ECG-ViT: A Transformer-Based ECG Classifier for Energy-Constraint Wearable Devices

The advancement in deep learning techniques has helped researchers acquire and process multimodal data signals from different healthcare domains. Now, the focus has shifted towards providing end-to-end solutions, i.e., processing these data and developing models that can be directly implemented on edge devices. To achieve this, the researchers try to solve two problems: (I) reduce the complex feature dependencies and (II) reduce the complexity of the deep learning model without compromising accuracy. In this paper, we focus on the later part of reducing the complexity of the model by using the knowledge distillation framework. We have introduced knowledge distillation on the Vision Transformer model to study the MIT-BIH Arrhythmia Database. A tenfold crossvalidation technique was used to validate the model, and we obtained a 99.7% F1 score and 99.3% accuracy. The model was further tested on the Xilinx Alveo U50 FPGA accelerator, and it is found fit for any low-powered wearable device implementation.

CONVERGENCE INVESTIGATION OF MULTIFRACTAL ANALYSIS BASED ON Lp-NORM CONSTRAINT

In this paper, we propose the multifractal detrending fluctuation analysis based on [Formula: see text]-norm constrain (MF-DFN). We assess the performance of the proposed algorithm by constructing a [Formula: see text]-model based multiplicative cascades time series. We calculate the generalized [Formula: see text] exponent [Formula: see text], [Formula: see text] exponent [Formula: see text], singularity exponent [Formula: see text] and multifractal spectrum [Formula: see text] for multifractal detrended fluctuation analysis (MF-DFA) and MF-DFN, respectively. We notice that under different norm constraints, the distribution of multifractal characteristics is quite different. Appropriate norm constraints can make the multifractal characteristics of time series described more accurately. Based on the analytical solution curve, the distribution of different multifractal numerical curves determines the correct selection of norm constraint value. In this study, using a combination of norms [Formula: see text] and [Formula: see text], the depicted numerical curves almost overlap the theoretical curve, which shows that the proposed MF-DFN is superior to MF-DFA. To eliminate the influence of specific time series, we reset parameters [Formula: see text] and [Formula: see text] in [Formula: see text]-model. Various experimental results also show the effectiveness of the proposed MF-DFN. In addition, we also verify the efficiency of the proposed MF-DFN by using the practical application. The experimental results show that our proposed method plays a significant role in ECG classification.