Wearable ECG signal processing for automated cardiac arrhythmia classification using CFASE‐based feature selection

Artificial Intelligence aims to mimic the human cognitive powers. AI is also into healthcare where it's been used for automating the diagnosis processes powered by increased availability of healthcare data. This research work develops an application, which can be used in diagnostic centers for multiple diseases diagnosis like Arrhythmia classification using ECG Image analysis, Malaria using Cell images, Brain tumor detection using MRI Scans and many more. At Present, this research work attempts to develop an Arrhythmia classification using ECG Image analysis and would add other diagnosis later on. Basically, ECG signals are used for analysis of Arrhythmia. The irregularities of heart beat are commonly called Arrhythmia. Here, a Convolutional Neural Network model is proposed for the classification of signals from ECG images to different classes. There are many different classes which are normal beat, Premature ventricular contraction beat, paced beat, atrial bundle branch block beat, left bundle branch block beat, atrial premature contraction beat, ventricular flutter wave beat, and ventricular escape beat, etc.

The electrocardiogram (ECG) is a diagnostic procedure that uses a skin electrode to record the heart's electrical activity. Heart diseases are the leading cause of mortality globally, and they have significant monetary cost. In the clinic, automated detection and classification technology of abnormalities in heart beats can assist physicians in making prompt and accurate medical diagnoses. This paper concentrates on two proposed models: Model A: 1-D 8-Layer CNN Model and Model B: 1- D CNN and the LSTM model to characterise ECG data into five categories: normal beat, right bundle branch block beat, left bundle branch block beat, premature ventricular contraction beat, and atrial premature beat. The data used to create and validate the models is obtained from the MIT-BIH Dataset, which comprises of 48 half-hour samples of two-lead continuous ECG signals obtained from 48 individuals. The dataset is divided into two files-.csv and.txt. For each sample, the.csv files include the readings collected from both leads. The experimental results show that Model A has an Accuracy = 99.68%, Precision = 99.23%, and a F1 score = 99.22%. And Model B has an Accuracy= 99.51%, Precision = 98.76%, and a F1 score = 98.76%. Our study aims to assist the medical sector by reducing the diagnoses time by automating the process of detecting arrythmias.

This paper presents a novel diagnosis method for the classification of cardiac arrhythmia based on electrocardiogram (ECG). The recognition of abnormal heart beat was performed using the context-independent hidden Markov model (HMM) and context-dependent HMM. Gaussian mixture model (GMM) was used as the state emission probability density function. The input feature of the recognition system was the raw heartbeat waveform. There were no other feature transform steps. Six classes of heartbeats, including normal beat, paced beat, left bundle branch block beat, right bundle branch block beat, premature ventricular contraction, and ventricular escape beat, were categorized by the system. The performance was evaluated on the MIT-BIH Arrhythmia Database and the European ST-T Database. The experimental results showed that the context-dependent HMM was more precise than the context-independent HMM. The context-dependent HMM system achieved a mean level of 95.26%, 90.67% and 95.60% in heart beats recognition accuracy, sensibility and specificity respectively. The results confirmed the importance of the generative model and the raw waveform feature strategy in ECG study.

Automatic ECG beat classification plays an important role in detecting cardiac disease. In this paper, we propose an automatic recognition model for ECG signals based on discrete wavelet transform (DWT), principal component analysis (PCA), kernel principal component analysis (KPCA), and adaptive kernel principal component analysis (AKPCA). We extracted different ECG features using DWT, PCA, KPCA, and AKPCA, respectively. These features were combined and used as support vector machine (SVM) input to classify the ECG. ECG records taken from the MIT-BIH arrhythmia database are selected to test the proposed method. The following five heartbeat types were classified using this method: normal beats (N), premature ventricular beats (V), right bundle branch block beats (R), left bundle branch block beats (L), and premature atrial beats (A). The sensitivity, accuracy, precision, and specificity reached 99.95%, 99.86%, 99.53%, and 99.70%, respectively. These results indicate the proposed method is reliable and efficient for ECG beat classification.

Detection and treatment of arrhythmias has become one of the main goals in cardiac care diagnosis provided by general practitioners. Electrocardiogram (ECG) analysis is one of the most commonly used tools to test and diagnose heart problems. Classification of ECG heartbeats enables the identification of specific arrhythmia or other heart conditions. This paper presents and contrasts the results from two effective ECG arrhythmia classification schemes. The first scheme consists of a principal component analysis (PCA) step for feature reduction at the input vector to the classifier, combined with soft independent modelling of class analogy (SIMCA). The second method uses a multi-class support vector machine (MSVM) classifier to differentiate between four different types of arrhythmia from ECG beats. The four types of beats include Normal (N), Premature Ventricular Contraction (PVC), and Atrial premature contraction (APC) and Right Bundle Branch Block Beat (RBBB). The time domain features were obtained from the St Petersburg INCART 12-lead Arrhythmia Database (incartdb). Between 10 and 30 Principal Components (PCs) were selected for reconstructing individual ECG beats and create the input vector to the classifier. The average classification accuracy of the proposed scheme is 76.83% and 98.33% using MSVM and SIMCA classifier respectively. The SIMCA classification algorithm provided better performance than the MSVM classifiers.

In this paper, QRS morphological features and the artificial neural network method was used for Electrocardiogram (ECG) pattern classification. Four types of ECG patterns were chosen from the MIT-BIH database to be recognized, including normal sinus rhythm, premature ventricular contraction, atrial premature beat and left bundle branch block beat. Authors propose a set of six ECG morphological features to reduce the feature vector size considerably and make the training process fast in addition to a simple but effective ECG heartbeat extraction scheme. Three types of artificial neural network models, MLP, RBF neural networks and SOFM were separately trained and tested for ECG pattern recognition and the experimental results of the different models have been compared. The MLP network exhibited the best performance and reached an overall test accuracy of 99.65%, and RBF and SOFM network both reached 99.1%. The performance of these classifiers was also evaluated in presence of additive Gaussian noise. MLP network was found to be more robust in this respect.

ECG beat classification using particle swarm optimization and radial basis function neural network

To assess the value of wearable 12-lead electrocardiogram (ECG) devices in pre-hospital diagnosis of acute ST segment elevation myocardial infarction (STEMI). This analysis was conducted among 441 patients selected from the''National ECG Network'', who used wearable 12-lead ECG device with critical situation warning of ST change between January 2019, and August, 2021.The general characteristics, response time and complaints of the patients with STEMI were analyzed.The accuracy of pre-hospital diagnosis of STEMI was compared between clinician's interpretation of ECGs and AI diagnosis by the wearable ECG device. In 89 of the patients, a pre-hospital diagnosis of STEMI was made by physicians based on ECGs from the wearable devices, and 58 of them sought medical attention after online warning, with a referral rate of 65.17%.The average time for diagnostic assessment of the ECGs was 153.02 s, and the average time for confirmation of the diagnosis was 178.06 s.The sensitivity for pre-hospital diagnosis of STEMI by clinician's interpretation of the ECGs and by AI diagnosis was 100% and 88.37%, respectively, with a specificity of 95.40% and 79.31%, respectively.The pre-hospital diagnosis by clinicians and AI diagnosis of STEMI both showed a high consistency with the subsequent definite clinical diagnosis of STEMI. Wearable 12-lead ECG devices can accurately record ECG characteristics of STEMI patients outside the hospital and allow immediate data uploading for an early diagnosis.The diagnoses of STEMI made based on AI technology are highly consistent with those by clinicians, demonstrating excellent clinical performance of the wearable ECG devices.

Background: Holter electrocardiogram (ECG) is the gold standard for ambulatory monitoring of atrial fibrillation (AF) but it is insufficient because of its limited recording time. Although several consumer ECG devices provide longer recording time, they generally do not undergo the regulatory process for medical use. Furthermore, current medical-grade devices for longer ECG monitoring are not continuous or too invasive for AF monitoring.A wearable ECG with a medical-grade dry textile electrode is a promising technology to remedy this limitation. This pilot study aimed to simultaneously compare the wearable and Holter ECGs for ambulatory monitoring in a clinical setting. Methods: This prospective observational study enrolled 18 patients who underwent AF ablation. One day after AF ablation, ambulatory ECG was obtained for three hours simultaneously using both the wearable and Holter ECG devices. Automatic ECG interpretations between devices were compared with correlation and agreement analyses. Results: Simultaneous ECG monitoring demonstrated a comparable analysis time and total heart beats between the two devices. Almost complete correlation and agreement were also demonstrated in all clinically relevant testing aspects except in R-wave amplitude (r = 0.743, p < .001). AF was detected in three patients. AF duration was the same in both ECG devices in two patients with continuous AF. In the remaining patient with intermittent AF, AF duration was shortened by 0.6% with the wearable ECG as compared to that with the Holter ECG. Conclusions: Simultaneous ECG comparison revealed a high consistency between the wearable and Holter ECG devices. The results of this study warrant further clinical studies for long-term monitoring of ambulatory ECG after AF ablation.

This paper presents a resource-saving system to extract a few important features of electrocardiogram (ECG) signals. In addition, real-time classifiers are proposed as well to classify different types of arrhythmias via these features. The proposed feature extraction system is based on two delta-sigma modulators adopting 250Hz sampling rate and three wave detection algorithms to analyze outputs of the modulators. It extracts essential details of each heartbeat, and the details are encoded into 68 bits data that is only 1.48% of the other comparable methods. To evaluate our classification, we use a novel patient-specific training protocol in conjunction with the MIT-BIH database and the recommendation of the AAMI to train the classifiers. The classifiers are random forests that are designed to recognize two major types of arrhythmias. They are supraventricular ectopic beats (SVEB) and ventricular ectopic beats (VEB). The performance of the arrhythmia classification reaches to the F1 scores of 81.05% for SVEB and 97.07% for VEB, which are also comparable to the state-of-the-art methods. The method provides a reliable and accurate approach to analyze ECG signals. Additionally, it also possesses time-efficient, low-complexity, and low-memory-usage advantages. Benefiting from these advantages, the method can be applied to practical ECG applications, especially wearable healthcare devices and implanted medical devices, for wave detection and arrhythmia classification.

심전도 신호 분석 및 부정맥 분류는 환자를 진단하고 치료하는데 중요한 역할을 한다. 부정맥은 맥박이 불규칙한 상태로 심실빈맥(VT)이나 심실세동(VF) 환자에게 심각한 위협이 될 수 있다. 심방조기수축(APC)과 상심실성빈맥(SVT), 심실조기수축(PVC)은 심실빈맥(VT)만큼 치명적이지는 않지만 심장질환을 진단하는데 중요한 부정맥이다. 본 논문은 2~3개의 부정맥 분류만을 고려한 기존의 방법을 극복하고 다양한 부정맥을 분류하기 위한 새로운 방법을 제시한다. 심전도 신호의 특징 추출을 위해서 EMD 방법으로 신호를 분해하여 IMFs를 얻는다. 입력 데이터의 양은 분류기 성능에 영향을 미치므로 신호 데이터의 차원을 감소시키기 위해 Burg 알고리즘을 IMFs에 적용하여 AR 계수를 구하고 여러 개의 이진 분류기를 결합한 다중 클래스 SVM의 입력으로 사용한다. 최적의 SVM 성능 파라미터를 선택하고 부정맥 분류에 적용한 결과 검출의 정확성은 96.8%~99.5%였다. 실험 결과는 제안한 EMD 방법에 의한 전처리 및 특징 추출과 다중 클래스 SVM에 의한 부정맥 분류의 유용성을 보여준다. Electrocardiogram(ECG) analysis and arrhythmia recognition are critical for diagnosis and treatment of ill patients. Cardiac arrhythmia is a condition in which heart beat may be irregular and presents a serious threat to the patient recovering from ventricular tachycardia (VT) and ventricular fibrillation (VF). Other arrhythmias like atrial premature contraction (APC), Premature ventricular contraction (PVC) and superventricular tachycardia (SVT) are important in diagnosing the heart diseases. This paper presented new method to classify various arrhythmias contrary to other techniques which are limited to only two or three arrhythmias. ECG is decomposed into Intrinsic Mode Functions (IMFs) by Empirical Mode Decomposition (EMD). Burg algorithm was performed on IMFs to obtain AR coefficients which can reduce the dimension of feature vector and utilized as Multi-class SVM inputs which is basically extended from binary SVM. We chose optimal parameters for SVM classifier, applied to arrhythmias classification and achieved the accuracies of detecting NSR, APC, PVC, SVT, VT and VP were 96.8% to 99.5%. The results showed that EMD was useful for the preprocessing and feature extraction and multi-class SVM for classification of cardiac arrhythmias, with high usefulness.

A previously reported method for electrocardiographic (ECG) telemetry in water using frequency-modulated current was improved to obtain more stable ECGs. The ECGs of seven healthy men were monitored using the improved method during and after whole-body submersion or underwater swimming. Bradycardia and arrhythmias were observed during the submersion, and transient tachycardia was detected after the start of underwater swimming, followed by bradycardia with arrhythmias. Three different types of arrhythmias were observed: sinus arrhythmia (SA), supraventricular extrasystole (SE) and ventricular extrasystole (VE). SA and SE tended to develop during the latter half of the period of submersion or underwater swimming, and especially after the restart of breathing. VEs were detected in only one subject during submersion, whereas they occurred in most subjects during and after underwater swimming. Individual variations were found in development of arrhythmias, one subject showing no arrhythmia. Bradycardia, SA and SE could depend on vagal suppression in underwater conditions, and VE may be related to the effect of muscular movement on cardiac function in addition to vagal inhibition.

Heart monitoring kits are only available for bedridden patients and the traditional heart monitoring kits have many wires that are obstacle patients’ mobility. Most of the existing heart monitoring kits can not detect heart diseases. Thus, the current study proposed a wireless heart monitoring kit to monitor patients with a heart abnormality. The proposed kit can detect and classify four arrhythmia types as well as normal ECG with high accuracy. The design and development of the wireless heart abnormality monitoring kit (WHAMK) in this research were divided into three stages. These stages are the development of an arrhythmias detection and classification method using artificial intelligence approach, design and implementation of the kit hardware, and design and coding of the kit software. Arrhythmias classification approach is divided into four stages, namely obtaining the electrocardiograph (ECG) signals, preprocessing, features extraction and classification. The features extraction method are based on statistical features. The library support vector machine (LIBSVM) was used to classify the ECG signals. The hardware of the kit is divided into two parts, namely ECG body sensor (EBS), and processing and displaying unit (PDU). EBS working on acquiring the ECG signal from patient's body. PDU working on processing the collected ECG signal, plotting it and detecting the arrhythmias. Arrhythmias classification approach was developed by using statistical features and LIBSVM. They were implemented in the software of the kit to enable it to detect the arrhythmias in the real-time and fully automatically. The kit can detect and classify four arrhythmia types as well as normal sinus rhythm (NSR). These types of arrhythmia are premature atrial contraction (PAC), premature ventricles contraction (PVC), Bradycardia and Tachycardia. The proposed kit gave a good accuracy for detecting and classifying Arrhythmia with the overall accuracy of 96.2%.

27 cardiac patients with different types of arrhythmias were treated in the coronary care unit with tiapamil administered by intravenous infusion. Special attention was given to determining the antiarrhythmic effects in patients with acute coronary insufficiency. In the group of 15 patients with ventricular extrasystoles (VEs; Lown classification III-V). 9 patients had acute myocardial infarction and 1 patient had coronary artery disease. In the group with supraventricular extrasystoles (6 cases). 1 patient had experienced anterior infarction. The third group comprised 6 patients with combined ventricular (Lown V) and supraventricular extrasystoles. In 1 case, the arrhythmia was due to acute anterior infarction. 2 patients had experienced reinfarction and one patient had crescendo angina pectoris. ECG and hemodynamics were monitored continuously before, during and for 20 h following therapy. In patients with VEs alone, the median frequency fell from 612 to 459/h at the third hour of infusion. The median VE/sinus beat quotient decreased from 0.125 to 0.0108 (p less than 0.01) in the third hour of treatment, and increased to 0.1029 after the completion of therapy. VE suppression was particularly marked in the 10 patients with coronary artery disease, 7 of these demonstrating a reduction by over 90%. Similar results were obtained in patients with supraventricular arrhythmias and mixed forms. The preliminary results show that tiapamil may be an effective antiarrhythmic agent in supraventricular and particularly in ventricular extrasystoles, and that its spectrum of antiarrhythmic action merits detailed investigation in larger numbers of patients.

In four patients with WPW syndrome atrial and ventricular premature beats were induced and the changes in form of the ventricular and atrial complexes were studied. Results indicate that, depending upon the timing of the premature atrial beat and the state of refractoriness of the His and Kent bundles, excitation of the ventricles occurs predominantly through the atrioventricular nodal system, predominantly through the Kent bundle or exclusively through one or both conduction systems. With short delays conduction through the Kent bundle may be blocked and only normal excitation of the ventricles occurs. In one patient with a history of attacks of tachycardia these normal QRS complexes were followed by retrograde activation of the atria by the Kent bundle, and attacks of supraventricular tachycardia of shorter or larger duration occurred. They stopped spontaneously, sometimes by delay or block, either of retrograde Kent conduction or of antegrade A-V nodal conduction, making it possible for the sinus node to capture the ventricles. They also could be terminated by induced atrial premature beats. In two patients tachycardias could be induced by appropriately timed ventricular premature beats during regular driving of the right ventricle. In one of these patients a circus movement, involving the Kent bundle, is probably present. By appropriate stimulation of the atria or ventricles during an attack of supraventricular tachycardia in this patient, one cycle length could be shortened without changing those of the following beats. These results suggest that a circus movement involving the atria, the normal atrioventricular conduction system and the Kent bundle is present. In the other patient, not fulfilling the WPW criteria, ventricular or atrial premature beats did not interfere with the basic rhythm of the tachycardia. Two hypotheses for this tachycardia are given: nodal tachycardia caused by rapid firing of the A-V node or a nodal tachycardia caused by a reciprocal mechanism in the A-V junction. The attacks could be blocked too by appropriately timed atrial and ventricular premature beats. No ventricular type of tachycardia could be demonstrated.