Search

An alarming increase in hypertension is a hazard to global health that poses severe implications for the body’s vital organs. To prevent serious repercussions, hypertension should be monitored continuously for early detection. It is well known that physiological signals, such as the photoplethysmogram (PPG) and electrocardiogram (ECG), carry essential information about the vitals of the human body. Considering this, numerous machine learning-based models based on ECG-PPG have been proposed for monitoring hypertension; however, such models are “non-explainable” and lack clinical interpretation. This work proposes a formal method-based runtime verification approach for hypertension monitoring using ECG and PPG sensing, which is explainable. The pulse arrival time (PAT) feature extracted using both signals is employed to implement a decision tree to infer hypertension patterns/policies defined in PAT, based on which a runtime monitor is synthesized to classify hypertension. Using the MIMIC II dataset, the proposed scheme’s performance is assessed, and the accuracy, sensitivity, and specificity are determined to be 95.7%, 93.9%, and 97.6%, respectively. [ABSTRACT FROM AUTHOR]

R peak detection is fundamental to the analysis of long-term electrocardiogram (ECG) signals. Despite their significant success in R peak detection, neural networks based on statistical learning usual require more than 50% of all data for training. However, it is often difficult to provide such a high proportion of training data in practice. This paper proposes a novel R peak detection method based on Generative Elastic Network (GEN), which is suitable for few-shot learning. Utilizing the Lobachevsky University Database (LUDB), this method achieves an accuracy exceeding 99% by using less than 3% of the data for training and 14% for validation. It dramatically reduces the dependency on large volumes of data for training and validation, while preserving an accuracy level that is on par with existing methods.

Advancements in integrated circuit (IC) technology have accelerated the miniaturization of body-worn sensors and systems, enabling long-term health monitoring. Wearable electrocardiogram (ECG), finger photoplethysmogram (PPG), and wrist-worn PPG have shown great success and significantly improved life quality. Chest-based PPG has the potential to extract multiple vital signs but requires ultra-high dynamic range (DR) IC to read out the small PPG signal among large respiration and artifacts inherent in daily life. This paper presents a dedicated high DR system for wearable chest PPG applications with a small form factor. The whole measurement system is integrated on a 20 cm2 PCB board. We have formulated a comprehensive evaluation protocol to validate the system with on-body chest PPG measurement in the workspace environment. First, chest PPG data was obtained from 6 adults and compared to data from a standard ECG patch. This system showed an average absolute deviation (AD) of 0.41 beats per minute, achieving > 99.53% heart rate (HR) accuracy. Second, chest PPG was recorded and compared to conventional PPG finger clip and PPG wristband, also showing > 98.6% HR matching and an absolute deviation in the standard deviation of NN intervals (SDNN) of < 12.8 ms for HRV monitoring within the protocol. Moreover, it successfully derives other vital parameters such as respiration rate and blood oxygen level (SpO2), showing the advancement among all these three reference modalities. This system can pave the way for new application areas, such as chest patches, to monitor chronic heart and respiratory diseases.

Coronary Artery Disease (CAD) is a key factor in several serious heart conditions, such as ischemic heart disease, myocardial infarction, and heart failure. Detecting and treating CAD early is crucial for preventing further progression of the disease. Computer-aided techniques are needed to automate the characterization of CAD conditions. In this paper, a novel approach for the automated identification of CAD utilizing improved energy estimation of electrocardiogram (ECG) signals is developed. Due to the multi-component nature of ECG and the nonlinear characteristics of energy operators, estimating their instantaneous energy may lead to cross-component interactions complicating the interpretation. To overcome this issue, variational mode decomposition with a data-driven and adaptively initialized number of modes is suggested to estimate the accurate energy of ECG beat. In this process, the instantaneous energy of each decomposed component is computed for enhanced energy estimation of each ECG beat. Multiple statistical features calculated from the localized region of ECG beat could effectively classify CAD-affected ECG beats with more than 99.80 % accuracy with an optimized ensemble classifier and a 10-fold cross validation scheme. Moreover, experimental findings have revealed the effectiveness of the proposed methodology when compared to contemporary approaches employing identical datasets.

Sleep apnea (SA) is one of the most prevalent sleep-related problems, impacting more than 100 million people worldwide. A full-night Polysomnography (PSG) is an effective SA diagnosis strategy. However, it requires multiple wearable devices and the patient staying overnight to collect the findings, rendering it both a time-consuming and costly option. Research attempts to develop non-invasive, sensor-based, or automated solutions for diagnosing SA are also made in recent years. In this study, we analyzed a total of 85 papers, shortlisted from an initial collection of 954 articles published in reputable scientific repositories, e.g., IEEE Xplore, PubMed Central (PMC), Springer, Elsevier etc., where each chosen study was thoroughly examined to determine its contribution and performance. A detailed analysis of data preprocessing, feature extraction and classification algorithm is also addressed. It is found that most of the studies are based on signal analysis for identifying sleep apnea, which yields results with substantial reliability, while contemporary research emphases have been on heart rate variability and pulse oximetry outcomes.

Wearable devices have gained significant popularity for continuous Electrocardiogram (ECG) monitoring due to their compactness and convenience. Day-by-day and 24/7 monitoring without gaps is demanded to promptly detect unusual symptoms that are short-term warning signs of dangerous diseases. Cost-saving criteria are the highest priority to extend the use time of devices and maintain the system operation. However, not all measured signals are useful because the signal quality will be affected by many things such as motion artifacts and muscle noise. Demands on classifying and sending only the usable signal is a requirement, it can save the battery on wearable devices and the expense on the server side (i.e., cloud-based processing), such as storage and processing resources. Therefore, Signal Quality Indices (SQIs) have been developed and researched to determine signal quality to meet the above requirements. This study introduces a novel SQI approach to classify signals. The proposed method has three contributions: 1) exponentially Weighted Mean-Variance (EWMV), a lightweight equation, to identify peaks, followed by applying an adaptive threshold to define peaks that have the same shape as R-peaks; 2) outlier elimination process is proposed to enhance the accuracy; and 3) maximal Overlap Discrete Wavelet Transform (MODWT) is employed to categorize ECG signals into a new class, potentially containing signals relevant to pathological analysis. Experimental results demonstrate that our algorithm achieves the highest sensitivity for both noisy and noiseless data sets. Specifically, it achieved a sensitivity of 99.31% for clean signals and 97.69% for noisy signals. In the case of the PhysioNet challenge data set, while our sensitivity of 96.37% may not be the highest, our accuracy stands out at 95.10%, surpassing other methods recently reviewed. Additionally, our approach demonstrates the lowest trade-off between sensitivity and accuracy among the surveyed SQI techniques.

Sudden cardiac death (SCD) is one of the main causes of death in athletes during exercise and physical activity. By analyzing the results of the ECG classification, doctors can promptly detect the presence of cardiovascular diseases and implement appropriate treatment measures, thereby reducing disease progression and the occurrence of complications. Previous studies have typically used machine learning methods that often required manual feature selection, which could be subjective, time-consuming, and limited in scope. Additionally, the relationships between features may be overlooked, leading to a decrease in model performance. Therefore, our proposed approach automatically learns and selects relevant features, avoiding the issues associated with manual feature selection. In this paper, we introduce the Grunet deep learning network model, which utilizes convolutional neural networks to automatically extract features from ECG signals, thereby enhancing feature utilization and representation capabilities. Given the time-sensitive nature of ECG signals, we introduce gated recurrent units (GRU) to better capture these features. The gating mechanism of GRU helps manage information flow and facilitates capturing long-term dependencies. In our experiments conducted on the MIT-BIH Arrhythmia Database and the European ST-T Database, our proposed method achieved an overall classification accuracy of 99.47% on the European ST-T Database, with a precision of 98.76% and a recall of 97.92% on the MIT-BIH Arrhythmia Database. The classification accuracy for classes N and Q reached 100%. The experimental results demonstrate that our method outperforms existing techniques, significantly reducing the cost of manual intervention and improving the accuracy of heartbeat classification.

Chest surface vibrations induced by cardiac activities can provide valuable insights into various heart conditions. Seismocardiogram (SCG) and Gyrocardiogram (GCG) signals, collectively referred to as Mechanocardiograms (MCG) and collected using a chest-mounted accelerometer and gyroscope, respectively, have the potential to serve as an effective alternative to Electrocardiograms (ECG) for continuous cardiac monitoring. In many cases, both modalities (MCG and ECG) can be used in tandem to monitor cardiac functions in both healthy subjects and Intensive Care Unit (ICU) patients. Direct acquisition of ECGs can be challenging in certain scenarios, such as with wearable devices, or due to issues with disconnections arising from loose contact surfaces or gel corrosion during long-term usage. ECG considered the gold standard for heart monitoring, is essential for a comprehensive assessment of cardiac parameters and patient health. MCGs have the potential to reliably estimate ECGs and can replace direct ECG acquisition procedures in such cases. In this study, we introduce M2ECG, a 1D-segmentation-based approach for translating ECG signals from the corresponding MCG signals acquired by an Inertial Measurement Unit (IMU) attached to the chest wall. Using the proposed SA-UNet, we achieved an average Pearson Correlation Coefficient (PCC) of 81.76% on a subject-independent test set. We also compared the estimated heart rates (HR) from the reconstructed ECGs to the ground truth ECGs to validate our model’s performance. The overall HR correlation achieved on the subject-independent test set was around 94.167%. The highest correlation of the HR and HRV calculated from the translated and the ground truth ECGs were around 99.073% and 96.289%, respectively for the best test case. The strong correlation observed in cardiac parameters (HR, HRV) underscores the effectiveness of MCG, suggesting its potential use for continuous monitoring of cardiac patients.

In this paper, we propose a data-shifting neural network (DSNN) for the detection of abnormal heartbeats. Our study aims to identify six types of electrocardiogram (ECG) signals using the deep learning network. In order to enhance the detection accuracy, the DSNN is devised by doubling the input signal using a data shifting scheme so that the amount of information for training may be adequately sufficient. Although the computational time doubles, the accuracy can be improved. When implemented using the Taiwan Semiconductor Manufacturing Company (TSMC) $0.18-\mu m$ complementary metal oxide semiconductor (CMOS) process, the proposed DSNN chip has an operating frequency at 20 MHz with chip area of $0.619 mm^{2}$ and maximum power dissipation $0.75 mW$ . As a result, the proposed DSNN can substantially increase detection accuracy for the task of ECG heartbeat classification. Results obtained after applying the proposed circuit to the ECG signals drawn from the MIT-BIH arrhythmia database showed that it achieved a detection rate of 97.17% with a small chip area, suggesting that it may be suitable for wearable or portable devices in healthcare.

Since the outbreak of coronavirus also known as COVID-19, there have been several studies on the disease. This study investigates patients’ electrocardiography (ECG) properties for an accurate prediction of this infectious disease. Our findings will be useful to medical practitioners in the accurate prognosis of COVID-19. We analyzed ECG datasets of patients who had tested positive for COVID-19 and Normal Persons who had tested negative. Using the analyzed dataset, we designed, developed, and evaluated twelve machine-learning models to predict COVID-19 with different combinations of learning algorithms and feature engineering techniques. Time, frequency, and time-frequency domain features were extracted using Time Series Feature Extraction Library (TSFEL). A combination of domains’ features comprising of Time, Frequency, and Time-Frequency attributes for prediction were also investigated. We deployed a T-test to determine the necessary and enough features for our predictive modeling. The ECG morphological feature analysis was based on peak-to-peak detection, onset, and offset of the QRS complex. K-nearest neighbors (KNN), Random Forest, and Support Vector Machine (SVM) learning algorithms were deployed to classify the ECG signals for the prediction of COVID-19. For each feature domain, the Shapley Additive Explanations (SHAP) visualization tool was used to ‘open’ the black box for explainability and interpretability of feature importance and weights. SHAP’s result suggests that extracted feature weight distribution in different algorithms shows consistency in feature importance per domain. However, the weight and the proportion are not the same. Performance evaluation of our classifiers was based on accuracy, sensitivity, precision, and F1 score. Our experimental results show that the classification encompassing all domain features (Temporal, Spectral, and wavelet) with Random Forest has a slight edge over other models. The proposed model has accuracy, sensitivity, precision, and F1 score of 97.09%, 98%, 97%, and 97%, respectively. COVID-19 continues to pose a threat to our world with its constant mutation, we believe that a study on its accurate prediction is critical for an effective universal mitigation strategy.

A large number of health and wellness applications wish to monitor electrocardiogram signal (ECG) in the least intrusive way possible. The use of dry textile electrodes with high input impedance bio-amplifiers is a common technique. This paper presents a novel ultra-high input impedance buffer to obtain a reduced input capacitance and enhance the acquisition of results for dry or non-contact capacitive ECG electrodes. Based on different bootstrapping techniques, input stage biasing and unipolar power supply, an operational amplifier buffer is implemented on Printed Circuit Board (PCB) to provide a low cost solution. The solution considers the protections for electrostatic discharges (ESD) and filters for interferences, maintaining high input impedance design. A method to measure input impedances in the range of femtofarads is also proposed. This active electrode prototype presents a 191 fF input capacitance at 50 Hz. Different skin-electrode impedance cases are studied to obtain the Common Mode Rejection Ratio (CMRR). As a result, a value of 81.4 dB is obtained with zero mismatch and 76.1 dB at 50 Hz is obtained with a resistive mismatch of 1–2.2 $\text{M}\Omega $ . Moreover, ECG test is evaluated using a discrete back-end circuit, digital filtering and software display. Finally, the full ECG circuit has been simulated under an ESD transient of ±8 kV IEC 61000–4-2 waveform. This paper demonstrated that the proposed solution with standard discrete components fulfills the industry requirements for robust ultra-high impedance ECG sensor.

Artificial intelligence (AI) is that encompasses machine learning (ML) combined with human intelligence had begun to reform medical practices into a new dimension. Advancements and developments of AI molds improved diagnostics in the field of cardiology. Electrocardiogram (ECG) is a simple and cost-effective tool to identify cardiac disorder and which is its reason for being into practice till date. Increasing the population of ECG big data annually requires automatic analysis and immediate interpretation for improved diagnosis. Modern AI techniques like deep learning (DL)-based convolutional neural networks (CNNs) provide an improved way of cardiac disease management and diagnosis. This review throws a light over application of AI in ECG analysis and its necessity. Rich sets of clinical ECG data curated carefully as private and public access developed for various cardiac and extra-cardiac diseases management. Rather than human ECG interpretation, AI can move modern medicine toward more personalized patient care. The intention of this review article is to assess clinical and research possibilities, gaps, and jeopardies involved in cardiac anomalies detection using ECG measurement. [ABSTRACT FROM AUTHOR]

Electrocardiography is the gold standard technique for detecting abnormal heart conditions. Automatic detection of electrocardiogram (ECG) abnormalities helps clinicians analyze the large amount of data produced daily by cardiac monitors. As thenumber of abnormal ECG samples with cardiologist-supplied labels required to train supervised machine learning models is limited, there is a growing need for unsupervised learning methods for ECG analysis. Unsupervised learning aims to partition ECG samples into distinct abnormality classes without cardiologist-supplied labels–a process referred to as ECG clustering. In addition to abnormality detection, ECG clustering has recently discovered inter and intra-individual patterns that reveal valuable information about the whole body and mind, such as emotions, mental disorders, and metabolic levels. ECG clustering can also resolve specific challenges facing supervised learning systems, such as the imbalanced data problem, and can enhance biometric systems. While several reviews exist on supervised ECG systems, a comprehensive review of unsupervised ECG analysis techniques is still lacking. This study reviews ECG clustering techniques developed mainly in the last decade. The focus will be on recent machine learning and deep learning algorithms and their practical applications. We critically review and compare these techniques, discuss their applications and limitations, and provide future research directions. This review provides further insights into ECG clustering and presents the necessary information required to adopt the appropriate algorithm for a specific application. [ABSTRACT FROM AUTHOR]

Usably secure ad-hoc device pairing fosters connectivity with hardware which is difficult to access (e.g., implanted) and grants convenience for ad-hoc short-term on-off pairing patterns (e.g. shared public domain). Examples are medical devices or fitness equipment. We present CardioID, an approach to extract features from heart rate variability for secure pairing keys that change with the randomness inherited in heart operation. Our processing chain is compatible with electrocardiogram (ECG, voltage), as well as ballistocardiogram (BCG, acceleration) type signals. Dissimilarities in locally generated sequences are accounted for using fuzzy cryptography exploiting Bose–Chaudhuri–Hocquenghem (BCH) codes. We propose a quantization to derive secure keys for cross BCG-ECG device pairing from heart-rate variability and analyze the performance in (inter- and intra-subject) BCG-to-ECG pairing. A secure communication protocol for Body Area Networks (BAN) is discussed. The attack surface of the protocol is analyzed, and we conduct a video-based attack study. In addition, two case studies with 5 (laboratory) and 20 (controlled in-field) subjects were conducted.

ECG (Electrocardiogram) data analysis is one of the most widely used and important tools in cardiology diagnostics. In recent years the development of advanced deep learning techniques and GPU hardware have made it possible to train neural network models that attain exceptionally high levels of accuracy in complex tasks such as heart disease diagnoses and treatments. We investigate the use of ECGs as biometrics in human identification systems by implementing state-of-the-art deep learning models. We train convolutional neural network models on approximately 81k patients from the US, Germany and China. Currently, this is the largest research project on ECG identification. Our models achieved an overall accuracy of 95.69%. Furthermore, we assessed the accuracy of our ECG identification model for distinct groups of patients with particular heart conditions and combinations of such conditions. For example, we observed that the identification accuracy was the highest (99.7%) for patients with both ST changes and supraventricular tachycardia. We also found that the identification rate was the lowest for patients diagnosed with both atrial fibrillation and complete right bundle branch block (49%). We discuss the implications of these findings regarding the reidentification risks of patients based on ECG data and how seemingly anonymized ECG datasets can cause privacy concerns for the patients.

Since stroke disease often causes death or serious disability, active primary prevention and early detection of prognostic symptoms are very important. Stroke diseases can be divided into ischemic stroke and hemorrhagic stroke, and they should be minimized by emergency treatment such as thrombolytic or coagulant administration by type. First, it is essential to detect in real time the precursor symptoms of stroke, which occur differently for each individual, and to provide professional treatment by a medical institution within the proper treatment window. However, prior studies have focused on developing acute treatment or clinical treatment guidelines after the onset of stroke rather than detecting the prognostic symptoms of stroke. In particular, in recent studies, image analysis such as magnetic resonance imaging (MRI) or computed tomography (CT) has mostly been used to detect and predict prognostic symptoms in stroke patients. Not only are these methodologies difficult to diagnose early in real-time, but they also have limitations in terms of a long test time and a high cost of testing. In this paper, we propose a system that can predict and semantically interpret stroke prognostic symptoms based on machine learning using the multi-modal bio-signals of electrocardiogram (ECG) and photoplethysmography (PPG) measured in real-time for the elderly. To predict stroke disease in real-time while walking, we designed and implemented a stroke disease prediction system with an ensemble structure that combines CNN and LSTM. The proposed system considers the convenience of wearing the bio-signal sensors for the elderly, and the bio-signals were collected at a sampling rate of 1,000Hz per second from the three electrodes of the ECG and the index finger for PPG while walking. According to the experimental results, C4.5 decision tree showed a prediction accuracy of 91.56% while RandomForest showed a prediction accuracy of 97.51% during walking by the elderly. In addition, the CNN-LSTM model using raw data of ECG and PPG showed satisfactory prediction accuracy of 99.15%. As a result, the real-time prediction of the elderly stroke patients simultaneously showed high prediction accuracy and performance.

Remote health monitoring has become a necessity due to reduced healthcare access resulting from pandemic lockdowns and the increasing aging population. Electrocardiography (ECG) is the standard for cardiac monitoring and arrhythmia identification, but it is inconvenient for long-time remote monitoring. Recently, Magnetocardiography (MCG) sensors that operate at room temperature became available based on spintronic sensors. However, MCG analysis is affected by the low-frequency noise present at the sensors. In this paper, we present an artificial intelligence (AI)-aided multi-model pipeline combining two AI architectures, defined as model-M1 and model-M2, targeted for ultra-edge Internet of Things (IoT) sensors to simulate arrhythmia detection. Model-M1 is a denoising preprocessor based on a sliding-window assisted deep-learning (DL) model. We investigate various methods to achieve high accuracy with lightweight computation. Model-M2 is a lightweight DL model that analyzes denoised ECG output from model-M1 to identify arrhythmia. We use multiple publicly available clinically annotated datasets to evaluate our proposal. We find that denoising by model-M1 retains the features, which assist the model-M2 in achieving high classification accuracy, compared to using a conventional moving average filter. This AI pipeline architecture is promising for privacy-preserving ultra-edge medical sensing devices.

Remote electrocardiogram (ECG) diagnosis with continuous real-time or near-real-time performance via a wireless wearable computing system would have significant value since it will enable on-time alerts and interventions, leading to life-saving outcomes. In this paper, a novel integration of open-loop and closed-loop switch modes is proposed, where the entire ECG diagnosis process is accomplished in three steps. First, the R-peak detection algorithm for initial diagnosis in local devices is executed, and 100% of the transmission energy is saved when no abnormality is detected. Second, in the case of an abnormality being detected, the edge device performs two-dimensional convolution neural network (2D-CNN) classification on the ECG signals, leading to either open-loop or closed-loop mode transmission based on the seriousness of the ECG signals. Third, in the cloud server, the received ECG signals may be further analyzed with a more sophisticated classification algorithm. The ECG classification accuracy ranges from 92.7 % to 99.1% depending on whether the analysis is executed locally (with reduced communication costs) or remotely (with increased computing resources). Overall, the ECG diagnosis process is partitioned into three components, including 1) irregular heartbeat detection, 2) 2-D CNN classification in the edge device, and 3) further classification in the cloud. The simulation results of such partition of ECG diagnosis in these layers supervised by the open-loop and closed-loop switch modes have demonstrated that the proposed system architecture can achieve efficient ECG diagnosis via wearable technologies with both reliable accuracies and reduced communication energy.

Electrocardiography (ECG) is generally deemed the golden standard for diagnosing cardiovascular diseases and photoplethysmography (PPG) is unobtrusive, low-cost, and convenient for continuous monitoring. However, PPG contains insufficient information to diagnose diseases. In this study, we propose a novel method to accurately convert PPG to ECG. The banded kernel ensemble method converts a low-quality source (PPG) to a high-quality destination (ECG). Unlike neural network solutions, our algorithm requires no computation burden in the conversion task after a trained model is obtained. The proposed algorithm is then tested on a publicly available MIMIC III database. Our prediction shows excellent accuracy in the validation dataset. It offers the testing performance of under 0.314 and above 0.55 in rrmse (relative root mean squared error) and KGE (Kling–Gupta efficiency), respectively, under the scenarios of three prevalent heart diseases. The reconstructed ECG can be further used to perform heart disease detection and we obtained an average correctness rate of 81%. Our method can help a large population of high-risk, believed-healthy persons to walk into doctors’ offices before the situation becomes irreversible.

For long-term electrocardiogram (ECG) signal monitoring, a portable and small size acquisition device with Bluetooth low energy (BLE) communication is designed and integrated with a Nvidia Jetson Xavier NX for realizing the electrode motion artifact removal technique. The digitalized ECG codes are converted from a front-end circuit, which contains several amplifiers and filters in the acquisition system. Thereafter, a zero padding scheme is applied for each 10-bits data to separate them into two-bytes data for BLE transmission. Xavier Edge AI platform receives these transmitted data and removes the electrode motion (EM) noise using the proposed low memory shortcut connection-based denoised autoencoder (LMSC-DAE). The simulation results demonstrate that the proposed algorithm significantly improves the signal-to-noise ratio (SNR) by 5.41 dB under the condition of SNRin = 12 dB, compared with convolutional denoising autoencoder with long short-term memory (CNN-LSTM-DAE) method. For practical test, an Arduino DUE platform is employed to generate noise interference by controlling a commercial digital-to-analog convertor. By combining the proposed ECG acquisition device with a non-inverting weighted summer, it can be applied to verify the reproducibility of measurement for the proposed method. The measurement results clearly indicate that the proposed LMSC-DAE has a higher improvement of SNR and lower percentage root-mean-square difference than the state-of-the-art Fully Convolutional Denoising Autoencoder (FCN-DAE).

STEMI is the most severe type of Myocardial Infarction that causes death or disability. Previous studies among physicians and paramedics have shown that the accuracy of STEMI diagnosis by the 12-lead ECG is not sufficient. Thus, we propose a 2D-CNN model that can detect and locate the STEMI signals from 12-lead ECG images. The 2D-CNN model is trained as a binary classification with 540 ECG images (270 STEMI cases and 270 other ECG images), and it achieved 96.3% accuracy, 96.2% sensitivity, 89.4% precision, 0.926 F1-score, and 0.962 ROC-AUC scores for 537 testing images. The proposed model is compared with 10 different transfer learning models. The proposed model has the best accuracy, sensitivity, F1-score, and ROC-AUC score. Grad-CAM technique is used for localization of STEMI signals in ECG images. According to the comparisons, the proposed model is the most reliable one for localizing the STEMI signals. This localization builds trust for the model because the CNN model is not a black box anymore, we can see where the CNN model looks and decides. The result of localization can also be used for teaching inexperienced physicians and paramedics. Also, the proposed model can be helpful for the accurate diagnosis of STEMI with a quick response time for clinical practices.

Remote Healthcare Monitoring Systems (RHMs) that use ECG signals are very effective tools for the early diagnosis of various heart conditions. However, these systems are still confronted with a problem that reduces their efficiency, such as energy consumption in wearable devices because they are battery-powered and have limited storage. This paper presents a novel algorithm for the compression of ECG signals to reduce energy consumption in RHMs. The proposed algorithm uses discrete Krawtchouk moments as a feature extractor to obtain features from the ECG signal. Then the accelerated Ant Lion Optimizer (AALO) selects the optimum features that achieve the best-reconstructed signal. Our proposed algorithm is extensively validated using two benchmark datasets: MIT-BIH arrhythmia and ECG-ID. The proposed algorithm provides the average values of compression ratio (CR), percent root mean square difference (PRD), signal to noise ratio (SNR), Peak Signal to noise ratio (PSNR), and quality score (QS) are 15.56, 0.69, 44.52, 49.04 and 23.92, respectively. The comparison demonstrates the advantages of the proposed compression algorithm on recent algorithms concerning the mentioned performance metrics. It also tested and compared against other existing algorithms concerning Processing Time, compression speed and computational efficiency. The obtained results show that the proposed algorithm extremely outperforms in terms of (Processing Time = 6.89 s), (compression speed = 4640.19 bps) and (computational efficiency = 2.95). The results also indicate that the proposed compression algorithm reduces energy consumption in a wearable device by decreasing the wake-up time by 3600 ms.

Electrocardiogram (ECG) acquisition is easily contaminated by interferences, and denoising is the most important task in ECG detection. The variational mode decomposition (VMD) algorithm is widely used in ECG denoising, which can overcome mode aliasing between intrinsic mode function (IMF) components that existed in the traditional empirical mode decomposition (EMD) algorithm, but the mode decomposition number ${K}$ and penalty factor $\alpha $ in VMD must be optimized to obtain the best signal decomposition accuracy. This article proposes an improved VMD denoising algorithm that overcomes the shortcomings of slow parameter selection and poor generalization in the traditional VMD algorithm. The algorithm presented first adopts the EMD algorithm to remove the low-frequency baseline drift noise and then employs the adaptive particle swarm optimization (APSO) algorithm to optimize the parameter pair (${K}$ , $\alpha $) for VMD. To validate the denoising performance of the improved VMD algorithm, the No. 103 record from the Massachusetts Institute of Technology (MIT) arrhythmia database is first selected as the pure ECG signal, then both 20-dB Gaussian white noises and 0.3-Hz baseline drift are added to simulate the noisy ECG signal. Second, the ECG signals of nine subjects are collected by a customized ECG detection platform based on AD8232 and ADALM1000. The ECG denoising results in simulation and actual experiments show that the improved VMD algorithm achieves the highest signal-to-noise ratio (SNR), correlation coefficient (CC), and minimum mean square error (MSE) compared with the traditional EMD and VMD algorithms, which demonstrates that the proposed denoising algorithm has stronger denoising ability and better retains morphological characteristics of the original ECG signals. [ABSTRACT FROM AUTHOR]

The Internet of Things (IoT)-enabled electrocardiogram (ECG) monitoring system is an evolution in healthcare, and it is used to monitor heart conditions during daily living conditions. However, the ECG signals from the wearable device are often contaminated with noise, such as power line interference, baseline wander, and muscle artefact. The presence of these noises increases the communication cost of IoT-enabled devices by transmitting unusable noisy signals as well as affecting the clinical decision-making at the application level. A real-time ECG noise detection system at the IoT-enabled gateway can reduce this communication cost and improve the overall quality of the transmitted signal. In this article, we propose a tunable ECG noise localization system to detect noisy ECG segments with a selected percentage of noise-free ECG levels. The proposed system will hold the following: 1) deployable ECG noise detector in the low-computational devices, such as smart phone; 2) reduce the data dropping rate at the network layer; 3) reduce the transmission cost; and 4) improve the reliability of the automated analysis of ECG at the IoT-enabled gateway. We evaluated the performance of the proposed ECG noise detector using publicly available and real-time ECG datasets from wearable ECG sensors. We also measured the data drop rate and compared it with the state-of-the-art algorithms. The impact of reduced data dropping rate also justifies using two commonly used R-peak detection algorithms. The proposed system reduces the data dropping rate by 21.09% on average over public and real-time datasets. Similarly, it increases the number of R-Peak detection by 15.33% on average compared with the existing binary noise classification system by saving partial noisy ECG segments. Thus, the proposed noise detection system improves the reliability of the automated ECG analysis. In summary, the study results support the necessity of a noise detection system for wearable ECG monitoring in daily living conditions to improve the data acquisition efficiency. In addition, a noise localizer can be used for dropping a noisy portion of the signal or filtering the only noisy segment of the ECG signal without blindly imposing a filter on both noise-free and noisy ECG signals. [ABSTRACT FROM AUTHOR]

At present, COVID-19 is still spreading and affecting millions of people worldwide. Minimizing the need for travel can significantly reduce the probability of infection and improve patients’ quality of life. The wireless body area network (WBAN) transmits the patients’ physiological data to the doctor remotely through the sensors in a way that minimizes physical contact with others. However, existing WBAN security authentication schemes have a core limitation that includes weak authentication performance and overconsumption of resources that precludes their widespread adoption in practical applications. Therefore, in this article, an enhanced dual-factor authentication system that addresses the mentioned drawbacks is proposed for securing WBAN resources. By combining iris and electrocardiogram (ECG) features, users would be required to pass the first-level iris authentication before performing the second-level ECG authentication, thus enhancing the overall security scheme of a WBAN system. Furthermore, we examined the existing inter-pulse-interval (IPI) encoding methods and propose a more efficient ECG IPI encoding algorithm, which can effectively shorten the encoding time without affecting the overall encoding performance. Finally, extensive experiments were performed to verify the performance of the proposed dual-factor iris and ECG-based WBAN authentication system using public iris and ECG databases. The experimental results show that the false acceptance rate (FAR) is close to 0.0%, and the false rejection rate (FRR) is close to 3.2%. Findings from this study suggest that the proposed dual-factor authentication scheme could aid the adequate deployment of security schemes to protect WBAN resources in practical applications. [ABSTRACT FROM AUTHOR]

Medical test selection is a recurring problem in health prevention and consists of proposing a set of tests to each subject for diagnosis and treatment of pathologies. The problem is characterized by the unknown risk probability distribution across the population and two contradictory objectives: minimizing the number of tests and giving the medical test to all at-risk populations. This article sets this problem in a general framework of chance-constrained medical test rationing with unknown subject distribution over an attribute space and unknown risk probability but with a given sample population. A new approach combining decision-tree and Bayesian inference is proposed to allocate relevant medical tests according to the subjects’ profile. Case studies on screening of hypertension and diabetes are conducted, and the performance of the proposed approach is evaluated. Significant savings on unnecessary tests are achieved with limited numbers of subjects needing but not receiving necessary tests. Note to Practitioners—Whether a medical test is needed for all subjects in health prevention? Is it possible to reduce unnecessary tests without jeopardizing the goal of screening at-risk populations? This article attempts to answer these questions by proposing a data-driven approach combining decision trees for subject profiling, Bayesian inference for unknown probability distribution estimation, and combinatorial optimization for test allocation. The application of this approach to a real-case study reduces the number of electrocardiogram (ECG) tests by 90% while keeping the number of hypertensive subjects needing but not receiving ECG tests small (five out of 230). A significant cut of unnecessary tests is also achieved in a second case study of diabetes screening. This approach allows decision-makers to better balance the cost-saving and the level of public health objective. Furthermore, the combination with decision trees makes the practical implementation quite straightforward. [ABSTRACT FROM AUTHOR]

In recent years, developments in wearable technologies and wireless communication have enabled wearable ECG monitoring. Long-term ECG monitoring result in large amounts of data. To sub-Nyquist sampling ECG signal, recently developed finite rate of innovation (FRI) technology is a practical option. However, existing FRI sampling schemes generally suffered from large model mismatch errors and insufficient reconstruction accuracy. In this paper, a sub-Nyquist sampling scheme based on the differentiated VPW optimization model is proposed for ECG signals sampling. The original ECG signal is modeled as a linear combination of several differentiated Variable Pulse Width (VPW) pulses and a model mismatch error signal. Then, a two-channel sampling framework is proposed to sample ECG signals with a sub-Nyquist rate. To improve the accuracy of the reconstruction, optimizing the reconstruction is considered. We formulate the objective function of the optimization to minimize the model mismatch error signal. A particle filter-based optimization method is proposed to solve the objective function with a high-dimensional variable. Block coordinate descent (BCD) technique is considered in the proposed optimization method. Finally, we conduct several simulations with real ECG signals from MIT-BIH Arrhythmia Database to test the performance of our proposed scheme. Simulation results have demonstrated the reconstruction accuracy of our proposed scheme outperforms other related schemes. [ABSTRACT FROM AUTHOR]

Objective: Based on cybernetics, a large system can be divided into subsystems, and the stability of each can determine the overall properties of the system. However, this stability analysis perspective has not yet been employed in electrocardiogram (ECG) signals. This is the first study to attempt to evaluate whether the stability of decomposed ECG subsystems can be analyzed in order to effectively investigate the overall performance of ECG signals, and aid in disease diagnosis. Methods: We used seven different cardiac pathologies (myocardial infarction, cardiomyopathy, bundle branch block, dysrhythmia, hypertrophy, myocarditis, and valvular heart disease) to illustrate our method. Dynamic mode decomposition (DMD) was first used to decompose ECG signals into dynamic modes (DMs) which can be regarded as ECG subsystems. Then, the features related to the DMs stabilities were extracted, and nine common classifiers were implemented for classification of these pathologies. Results: Most features were significant for differentiating the above-mentioned groups (p value<0.05 after Bonferroni correction). In addition, our method outperformed all existing methods for cardiac pathology classification. Conclusion: We have provided a new spatial and temporal decomposition method, namely DMD, to study ECG signals. Significance: Our method can reveal new cardiac mechanisms, which can contribute to the comprehensive understanding of its underlying mechanisms and disease diagnosis, and thus, can be widely used for ECG signal analysis in the future. [ABSTRACT FROM AUTHOR]

This paper adopts a holistic approach to stress detection issues in software and hardware phases and aims to develop and evaluate a specific low-power and low-cost sensor using physiological signals. First, a stress detection model is presented using a public data set, where four types of signals, temperature, respiration, electrocardiogram (ECG), and electrodermal activity (EDA), are processed to extract 65 features. Using Kruskal-Wallis analysis, it is shown that 43 out of 65 features demonstrate a significant difference between stress and relaxed states. K nearest neighbor (KNN) algorithm is implemented to distinguish these states, which yields a classification accuracy of 96.0 ± 2.4%. It is concluded that ECG and EDA signals are sufficient to detect stress with high accuracy while the system has fewer sensors and consumes lower power. Second, a novel sensor is developed to collect ECG and EDA signals from eighteen healthy participants aged 16 to 40 years. They were exposed to stress during an arithmetic mental task and Stroop Color-Word Test. This sensor has a battery life of 70 hours and can detect stress with 94.4 ± 2.5% accuracy. The proposed software and hardware can potentially be used in a wearable device for continuous stress detection. [ABSTRACT FROM AUTHOR]

In this letter, for the first time, we propose a unified framework based on attentive cycle-generative adversarial network for the synthesis of electrocardiogram (ECG) signals from the seismocardiogram (SCG) signals. The proposed attentive cycle generative adversarial network exploits dual generators and dual discriminators to learn the pattern for the synthesis of ECG from SCG and vice versa. The proposed framework is evaluated on publicly available combined measurement of ECG, breathing and seismocardiogram (CEBS) database. Subjective visual analysis and objective performance metrics demonstrate that the proposed framework can accurately derive the ECG signal from SCG signal. Since, the SCG can be recorded using a wearable and non-adhesive modality, it can provide comfort to the patients by avoiding adhesive ECG electrodes. Further, the derived ECG can help in better cardiac rhythm and arrhythmia analysis. [ABSTRACT FROM AUTHOR]

Breathing Rate (BR) is a key physiological parameter measured in a wide range of clinical settings. However, it is still widely measured manually. In this paper, a novel framework is proposed to estimate the BR from an electrocardiogram (ECG), a photoplethysmogram (PPG), or a blood pressure (BP) signal. The framework uses Empirical Mode Decomposition (EMD) and Discrete Wavelet Transform (DWT) methods to extract respiratory signals, taking advantage of both time and frequency domain information. An Extended Kalman Filter (EKF), incorporating a Signal Quality Index (SQI), enabled our method to achieve acceptable performance even for significantly distorted periods of the signals. Using state vector fusion, the output signals are combined and finally the BR is estimated. The framework was tested on two publicly available clinical databases: the MIT-BIH Polysomnographic and BIDMC databases. Performance was evaluated using the mean absolute percentage error (MAPE). The results indicated high accuracy: MAPEs on the two databases of 3.9% and 3.6% for ECG signals, 6.0% for PPG, and 5.0% for BP signals. The results also indicated high robustness to noise down to $0dB$ . Therefore, this framework may have utility for BR monitoring in high noise settings.

Due to the proliferation of the Internet of Things (IoT), the IoT devices are becoming utilized at the edge network at a much higher rate. Conventionally, the IoT devices lack the computation resources required for carrying out ultra-edge analytics. In this paper, we go beyond the typical edge analytics paradigm, which is mostly limited to user-smartphones, and investigate how to embed intelligence into the ultra-edge IoT sensors. To conceptualize the smart IoT sensors with enhanced intelligence, we select the arrhythmia detection task employing Electrocardiogram (ECG) trace as one of the mobile health (mHealth) cases. The existing approaches are not feasible for ultra-edge IoT sensors due to the extensive noise-filtering and manual feature extraction phase. Hence, in this paper, to facilitate the analytics, we propose a Deep Learning-based Lightweight Arrhythmia Classification (DL-LAC) method, which employs only single-lead ECG trace and does not require noise-filtering and manual feature extraction steps. As the proposed technique, we design a one-dimensional Convolutional Neural Network (CNN) architecture. Complying with the ANSI/AAMI EC57:1998 standard, four heartbeat types are taken into consideration as class labels. The efficiency and the generalization ability of the proposed model are evaluated, employing four different datasets from PhysioNet. The experimental results demonstrate that the proposed DL method outperforms traditional methods such as the Delay Differential Equation (DDE)-based optimization, K-Nearest Neighbor (KNN), and Random Forest (RF). The proposed DL-LAC illustrates encouraging performance in terms of time and memory requirement when the trained model is transferred to virtualized microcontrollers connected to IoT sensors.

Recent innovations in wearable electrocardiogram (ECG) devices have enabled various personal healthcare applications based on heart rate variability (HRV). However, wearable ECGs rarely undergo visual inspection by medical experts, hence may contain noise and artifacts. Because apparent changes in the recorded ECGs caused by noise and artifacts may hamper the extraction of QRS complexes, an R-R interval (RRI) estimation algorithm tolerant to these measurement faults is required as the initial step toward HRV analysis using wearable ECGs. This paper proposes a semi-real-time RRI estimation for wearable ECGs utilizing a two-stage structure. In the preprocessing stage, we use a complex-valued wavelet that can adaptively fit to morphological variations of the QRS complex while retaining computing resources for extracting the QRS complex features. In the decision stage, we make use of complex-valued features and select appropriate QRS complexes in consideration of three features: peak magnitude, peak location, and peak morphology (phase). Initial evaluations show that the QRS complex detection performance of the proposed method achieved the F1 score of 0.952 ± 0.040 when targeting pseudo ECG data created from open data assuming wearable ECGs, and of 0.986 ± 0.018 when targeting actual ECG data recorded by a shirt-type wearable ECG device during an exercise activity. Furthermore, the proposed method was able to suppress overlook or misdetection of QRS complexes, so the obtained RRIs are closer to the reference RRIs. The proposed method therefore contributes to achieving accurate HRV analysis using wearable ECGs in terms of obtaining accurate RRIs.

Gelatine based phantoms for electrophysiology are becoming widely used as they allow the controlled validation of new electrode and new instrumentation designs. The phantoms mimic the electrical properties of the human body and allow a pre-recorded electrophysiology signal to be played-out, giving a known signal for the novel electrode or instrumentation to collect. Such controlled testing is not possible with on-person experiments where the signal to be recorded is intrinsically unknown. However, despite the rising interest in gelatine based phantoms there is relatively little public information about their electrical properties and accuracy, how these vary with phantom formulation, and across both frequency and duration of use. This paper investigates ten different phantom configurations, characterising the impedance of the gelatine and electrodes, comparing this to both previously reported electrical models of Ag/AgCl electrodes placed on human skin and to a model made from ex vivo porcine skin. This article shows how the electrical properties of the phantoms can be tuned using different concentrations of gelatine and of sodium chloride (NaCl) added to the mixture, and how these properties vary over the course of seven days for a.c. frequencies in the range 20–1000 Hz. The results demonstrate that gelatine phantoms can accurately mimic the frequency response properties of the body–electrode system to allow for the controlled testing of new electrode and instrumentation designs.

An automatic system for heart arrhythmia classification can perform a substantial role in managing and treating cardiovascular diseases. In this paper, a deep learning-based multi-model system is proposed for the classification of electrocardiogram (ECG) signals. Two different deep learning bagging models are introduced to classify heartbeats into different arrhythmias types. The first model (CNN-LSTM) is based on a combination of a convolutional neural network (CNN) and long short-term memory (LSTM) network to capture local features and temporal dynamics in the ECG data. The second model (RRHOS-LSTM) integrates some classical features, i.e. RR intervals and higher-order statistics (HOS), with LSTM model to effectively highlight abnormality heartbeats classes. We create a bagging model from the CNN-LSTM and RRHOS-LSTM networks by training each model on a different sub-sampling dataset to handle the high imbalance distribution of arrhythmias classes in the ECG data. Each model is also trained using a weighted loss function to provide high weight for not sufficiently represented classes. These models are then combined using a meta-classifier to form a strong coherent model. The meta-classifier is a feedforward fully connected neural network that takes the different predictions of bagging models as an input and combines them into a final prediction. The result of the meta-classifier is then verified by another CNN-LSTM model to decrease the false positive of the overall system. The experimental results are acquired by evaluating the proposed method on ECG data from the MIT-BIH arrhythmia database. The proposed method achieves an overall accuracy of 95.81% in the “subject-oriented” patient independent evaluation scheme. The averages of F1 score and positive predictive value are higher than all other methods by more than 3% and 8% respectively. The experimental results show the superiority of the proposed method for ECG heartbeats classification compared to many state-of-the-art methods.

Premature ventricular contraction (PVC) is one of the most common arrhythmias which can cause palpitation, cardiac arrest, and other symptoms affecting the work and rest activities of a patient. However, patients hardly decipher their own feelings to determine the severity of the disease thus, requiring a professional medical diagnosis. This study proposes a novel method based on image processing and convolutional neural network (CNN) to extract electrocardiography (ECG) curves from scanned ECG images derived from clinical ECG reports, and segment and classify heartbeats in the absence of a digital ECG data. The ECG curve is extracted using a comprehensive algorithm that combines the OTSU algorithm with erosion and dilation. This algorithm can efficiently and accurately separate the ECG curve from the ECG background grid. The performance of the classification model was evaluated and optimized using hundreds of clinical ECG data collected from Fujian Provincial Hospital. Additionally, thousands of clinical ECG reports were scanned to digital images as the test set to confirm the accuracy of the algorithm for practical application. Results showed that the average sensitivity, specificity, positive predictive value, and accuracy of the proposed model on the MIT-BIH dataset were 95.47%, 97.72%, 98.75%, and 98.25%, respectively. The classification average sensitivity, specificity, positive predictive value, and accuracy based on clinical scanned ECG images can reach to 97.24%, 81.6%, 83.8%, and 89.33%, respectively, and the clinical feasibility is high. Overall, the proposed method can extract ECG curves from scanned ECG images efficiently and accurately. Furthermore, it performs well on heartbeat classification of normal (N) and ventricular premature heartbeat.

Objective: Noise and disturbances hinder effective interpretation of recorded ECG. To identify the clean parts of a recording, free from such disturbances, various quality indicators have been developed. Previous instances of these indicators focus on human-defined desirable properties of a clean signal. The reliance on human-specified properties places an inherent limitation on the potential power of signal quality indicators. To move away from this limitation, we propose a data-driven quality indicator. Methods: We use an unsupervised deep learning model, the auto-encoder, to derive the quality indicator. For different quality assessment settings we compare the performance of our quality indicator with traditional indicators. Results: The data-driven method performs consistently strong across tasks while performance of the traditional indicators varies strongly from task to task. Conclusion: This strong performance indicates the potential of data-driven quality indicators for use in ECG processing, removing the reliance on expert-specified desirable properties. Significance: The proposed methodology can easily be extended towards learning quality indicators in other data modalities. [ABSTRACT FROM AUTHOR]

This paper proposes a novel deep learning architecture involving combinations of Convolutional Neural Networks (CNN) layers and Recurrent neural networks (RNN) layers that can be used to perform segmentation and classification of 5 cardiac rhythms based on ECG recordings. The algorithm is developed in a sequence to sequence setting where the input is a sequence of five second ECG signal sliding windows and the output is a sequence of cardiac rhythm labels. The novel architecture processes as input both the spectrograms of the ECG signal as well as the heartbeats’ signal waveform. Additionally, we are able to train the model in the presence of label noise. The model’s performance and generalizability is verified on an external database different from the one we used to train. Experimental result shows this approach can achieve an average F1 scores of 0.89 (averaged across 5 classes). The proposed model also achieves comparable classification performance to existing state-of-the-art approach with considerably less number of training parameters. [ABSTRACT FROM AUTHOR]

Epilepsy is one of the most prevalent paroxystic neurological disorders. It is characterized by the occurrence of spontaneous seizures. About 1 out of 3 patients have drug-resistant epilepsy, thus their seizures cannot be controlled by medication. Automatic detection of epileptic seizures can substantially improve the patient’s quality of life. To achieve a high-quality model, we have to collect data from various patients in a central server. However, sending the patient’s raw data to this central server puts patient privacy at risk and consumes a significant amount of energy. To address these challenges, in this work, we have designed and evaluated a standard federated learning framework in the context of epileptic seizure detection using a deep learning-based approach, which operates across a cluster of machines. We evaluated the accuracy and performance of our proposed approach on the NVIDIA Jetson Nano Developer Kit based on the EPILEPSIAE database, which is one of the largest public epilepsy datasets for seizure detection. Our proposed framework achieved a sensitivity of 81.25%, a specificity of 82.00%, and a geometric mean of 81.62%. It can be implemented on embedded platforms that complete the entire training process in 1.86 hours using 344.34 mAh energy on a single battery charge. We also studied a personalized variant of the federated learning, where each machine is responsible for training a deep neural network (DNN) to learn the discriminative electrocardiography (ECG) features of the epileptic seizures of the specific person monitored based on its local data. In this context, the DNN benefitted from a well-trained model without sharing the patient’s raw data with a server or a central cloud repository. We observe in our results that personalized federated learning provides an increase in all the performance metric, with a sensitivity of 90.24%, a specificity of 91.58%, and a geometric mean of 90.90%. [ABSTRACT FROM AUTHOR]

This review investigated research works on affective computing by using electrocardiogram (ECG) and electrodermal activity (EDA). The 27 related research papers, including 23 from IEEE Journals and 4 from other Q1 Journals in the last five years, were studied. The main goals have been to summarize common trends in this field in recent years as well as provide discussions and a big picture of how future research should be done. The content of this review covers the fundamental pipeline of affective computing research consisting of stimuli, intelligent affective computing, and sensor design and data processing. Additionally, we discuss future research directions and potential applications, including stress monitoring, music cognition, and robotics, where ECG and EDA will play a significant role with wearable devices. Therefore, this review serves as an information summary for everyone who is interested in affective computing and improving AI’s understanding of human emotion. [ABSTRACT FROM AUTHOR]

Computer-aided detection and diagnosis in ECG signals for heart diseases are gaining increasing attention. However, developing and selecting the highly performing diagnostic model suitable for clinical implications is still challenging. In this paper, we proposed a combined network of convolutional neural network (CNN) and Recurrent Neural Network (RNN), designed for the classification of ECG heart signals for diagnostic purpose. The proposed network consists of 2 convolutional layers with 5 × 5 kernels and ReLU activations, followed by 4 residual blocks, 2 bidirectional long short-term memory (biLSTM) layers, as well as 2 fully connected layers. Each residual block involved the structure of a Squeeze-and-Excitation Network (SENet) with lightweight property to recalibrate the feature map of the network. The last dense layer has 5 outputs, equivalent to the classes considered: non-ectopic beats, supraventricular ectopic beats, ventricular ectopic beats, fusion beats, and unknown beats. To train and evaluate the combined CNN and RNN, we transferred the knowledge acquired on beat classification tasks in 2017 PhysioNet/CinC Challenge to that in PhysionNet's MIT-BIH dataset. The developed network achieved a recognition sensitivity of 95.90%, accuracy = 95.90% and specificity = 96.34% with classification time of single sample = 6.23 s in detecting 5 ECG classes. A comparative analysis proved the high performance of the proposed combined CNN and RNN against previous methods, demonstrating the potential of our proposed network in the analysis of beat patterns. The proposed model can be applied in cloud computing or implemented in mobile devices to evaluate cardiac health with the highest precision.

Compressed Sensing (CS) has been proposed as a low-complexity ECG data compression scheme for wearable wireless bio-sensor devices. However, CS decoding is characterized by high computational complexity. As a result, it represents a burden to the computational and energy resources of the network gateway node, where decoding is performed. In this article, we propose a Fast Compressive Electrocardiography (FCE) technique to address this problem. CS decoding in FCE is based on Weighted Regularized Least-Squares (WRLS), rather than the standard approach based on ℓ1 norm minimization. The WRLS formulation takes into account prior knowledge of ECG signal properties to estimate an optimally compact and accurate representation of ECG signals. Numerical results show that decoding by FCE is on average 33 times faster than the fastest tested CS-based ECG decoding technique. In addition, high-quality ECG signal reconstruction by FCE is achieved at 32% higher compression ratio. Therefore, FCE can contribute to improving the overall energy and computational resource efficiency of CS-based remote ECG monitoring systems.

In this paper, we propose a novel machine learning pipeline to detect QRS complexes in very noisy wearable electrocardiogram (ECG) devices. The machine learning pipeline consists of a Butterworth filter, two wavelet convolutional neural networks (WaveletCNNs) autoencoders, an optional QRS complex inverter, a Monte Carlo k-nearest neighbours (k-NN), and a convolutional long short-term memory (ConvLSTM). WaveletCNN autoencoders filter out electrode contact noise, instrumentation noise, and motion artifact noise by using the advantages of wavelet filters and convolutional neural networks. The QRS complex inverter flips inverted QRS complexes. Monte Carlo k-NN performs automatic gain control on the ECG signals in order to normalize it. The ConvLSTM executes the final QRS complex detection by using the power of a convolutional neural network and a long short-term memory. The MIT-BIH, the European ST-T, and the Long Term ST database Noise Stress Test databases provide the training and testing ECG recordings. The proposed machine learning pipeline performs 3 standard deviations better than the state of the art QRS complex detection algorithms in terms of F1 score for very noisy environments.

Heart attack else wise termed as myocardial infarction (MI) causes irreparable death of cardiac muscles yielding the focal reason for most casualties among all cardiovascular diseases (CVDs’). A 12-lead electrocardiogram (ECG) generally depicts cardiac abnormalities and so customary deep learning (DL) approaches use the whole signal for binary detection purposes, that is separating healthy control (HC), and MI classes. This paper proposes an alternative approach where 21 temporal features in lieu of the temporal signal are collected from the 12 lead data to reduce redundancy and class imbalance keeping the vital information intact. Then these extracted features are fed into a detection model consisting of a one dimensional (1-D) convolutional neural network (CNN) and a bidirectional long short-term memory (bi-LSTM) layer which classifies into three classes, namely: HC, MI, and non-myocardial infarction (non-MI) subjects for a realistic and reliable assessment. The model’s performance is evaluated using 517 records acquired from the Physikalisch-Technische Bundesanstalt (PTB) database and a state-of-art overall accuracy of 99.246%, kappa of 0.983, and macro averaged F1 score of 98.86% were achieved using stratified 5-fold cross-validation. DL methods suffer to make unbiased decisions in the case of class imbalance due to an insufficient amount of data for a particular class and thus temporal features are employed to inherently reduce this problem. The successful performance of the extracted features depends on the precise detection of fiducial points, and so multiple novel algorithms have been introduced in this paper. [ABSTRACT FROM AUTHOR]

Autonomic nervous system (ANS) can maintain homeostasis through the coordination of different organs including heart. The change of blood glucose (BG) level can stimulate the ANS, which will lead to the variation of Electrocardiogram (ECG). Considering that the monitoring of different BG ranges is significant for diabetes care, in this paper, an ECG-based technique was proposed to achieve non-invasive monitoring with three BG ranges: low glucose level, moderate glucose level, and high glucose level. For this purpose, multiple experiments that included fasting tests and oral glucose tolerance tests were conducted, and the ECG signals from 21 adults were recorded continuously. Furthermore, an approach of fusing density-based spatial clustering of applications with noise and convolution neural networks (DBSCAN-CNN) was presented for ECG preprocessing of outliers and classification of BG ranges based ECG. Also, ECG's important information, which was related to different BG ranges, was graphically visualized. The result showed that the percentages of accurate classification were 87.94% in low glucose level, 69.36% in moderate glucose level, and 86.39% in high glucose level. Moreover, the visualization results revealed that the highlights of ECG for the different BG ranges were different. In addition, the sensitivity of prediabetes/diabetes screening based on ECG was up to 98.48%, and the specificity was 76.75%. Therefore, we conclude that the proposed approach for BG range monitoring and prediabetes/diabetes screening has potentials in practical applications. [ABSTRACT FROM AUTHOR]

An ultra-low-power low-noise analog front end (AFE) is presented in this work, aiming for long-term ECG with clear P-waves for clinical diagnose. The chopper amplifier with passive noise filter and PWM based offset cancellation results in an input-reference noise of 0.39 μVrms, which shows 3.7X noise improvements among the state-of-the-art designs. With the digital offset cancellation by the pulse-width modulation wave, the AFE achieves a low input-referred dc offset of 0.4 μV among 9 tested chips and a low drift under 30 μV. A dynamic scale ADC with low-power comparator strategy prevents the instability and signal-loss, achieving an SFDR of 71.6 dB. The proposed AFE achieves a noise-efficient-factor (NEF) of 2.4 with a power consumption of 4 μW. The fabricated chip is demonstrated in a miniature prototype for long-term ECG monitoring application, presenting a clear ECG waveform with visible P-wave. The simultaneously ECG recording with a medical grade 12-lead ECG Holter shows the effective acquisition of the prototype, proofing the better noise performance. [ABSTRACT FROM AUTHOR]

Long-term electrocardiogram (ECG) signal monitoring necessitates a large amount of memory space for storage, which affects the transmission channel efficiency during real-time data transfer. Using a combination of tunable-Q wavelet transform (TQWT) and adaptive Fourier decomposition (AFD), the proposed work develops a new single-channel ECG signal compression algorithm. The input parameters of TQWT were selected so that the lowest frequency subband contained highest energy along with minimal loss. A new Möbius transform-based AFD was introduced to improve the fidelity, by computing highest energy coefficients using Nevanlinna factorization, with suitable decomposition level. Finally, the lossless compression was performed in polar coordinate of final complex coefficients that significantly improved the compression ratio (CR). The algorithm was tested on “python” programming platform, tested in Raspberry Pi (R-Pi), for real-time data processing, and wireless transmission to cloud server and smartphone devices. The suggested work yielded CR, percent root mean square error (PRD), and PRD normalized (PRDN) of 30.06, 7.80, and 11.62, respectively, after testing on 48 Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) ECG data with 30-min duration. A rigorous quality assessment of the reconstructed signal ensured that there was minimal impact on various characteristic domains in the ECG signal, enhancing its acceptability in medical applications. [ABSTRACT FROM AUTHOR]

Electrocardiogram (ECG) biometric recognition as a personal identification method is receiving more and more attention because it can support live verification results. Compared with other biometric-based methods, it can provide higher security performance. The difficulty of the problem lies in how to stably extract ECG signal features and achieve real-time verification. In this study, a new type of sparse representation learning framework called statistical $n$ -best adaptive Fourier decomposition (SAFD) originated by Qian is adopted in ECG biometric identification. Adaptive Fourier decomposition (AFD) is a recently developed combination of transform-based signal decomposition and sparse representation method, which can adaptively select the atoms from a redundant dictionary through orthogonal processing. The advantage of the AFD-type methods is that each atom in the dictionary has a precise mathematical formula with good analytic properties. This characteristic is significantly distinguished it from other existing sparse representations, where the atoms learned are usually matrix data and cannot be described mathematically. The proposed SAFD extends the existing $n$ -best AFD from processing single signal to multi-signals and implements the $n$ -best AFD in the stochastic Hardy space. Therefore, the small number of learned atoms by SAFD is sufficient to capture internal structure and robustness of the signal and generate a discriminative representation that reflects the time–frequency characteristics of signals. It is very suitable for non-stationary signals like ECG. The proof of convergence of the algorithm is presented. Extensive experiments are conducted on five public databases collected in different realistic conditions, and an average identification accuracy of 98.0% is achieved. In addition, less than 1 ms for one matching process makes it possible to be implemented in real time. Experimental results demonstrate that the proposed method can achieve superior performance compared to other state-of-the-art ECG biometric identification methods. [ABSTRACT FROM AUTHOR]

Electrocardiogram (ECG) has attracted intense research interests and contests due to its great value in biometric applications, especially with the technological progress of smart instrumentation and artificial intelligence nowadays. For the issue of ECG-based human identification, distance measure in a suitable data space can be an effective solution. Almost all conventional distance measures rely on the independent data samples for classification. However, such measuring mechanism is vulnerable to the variation and bias of data distribution, which are easily caused by the noisy artifacts in this issue. To tackle this problem, we suggest doing distance measurement at the level of multiple-set bundle that consists of multiple sample sets obtained under different conditions. More specifically, we propose a novel method, “Collaborative-Set Measurement (CSM),” which creatively extends distance measure from the sample level through the set level to the bundle level. CSM not only captures the representative information from within-set distribution but also exploits the discriminative information from between-set collaboration, and meanwhile incorporates the robust information from in-bundle fusion. Attributing to these information, this method can largely enhance the identification performance in this issue despite the serious noisy artifacts influencing the data samples. The experimental results have demonstrated the capability of CSM for ECG-based human identification, by means of the public challenging database, DREAMER, on which the proposed method produces an accuracy of 91.30%. In summary, this proposal has initially brought forward a new thought of “collaborative sets.” This thought may provide inspirations for further researches on biometric recognition and general signal classification topics. [ABSTRACT FROM AUTHOR]

Electrocardiogram (ECG) is a tool to help judge heart activity. In recent years, the convolutional neural network (CNN) and various deep learning algorithms have been widely used in ECG diagnosis. CNN only considers the local feature. However, the ECG signal is susceptible to noise, and the waveform is complex, making it difficult for existing methods to get a good result. This article presents a novel neural network architecture for ECG diagnosis based on deformable CNN (Deform-CNN). The architecture makes good use of the feature-learning capability of deformable convolution to learn the time-domain and lead characteristics of multilead ECG signals. The proposed end-to-end method can achieve an overall diagnostic accuracy of 86.3% in the 12-lead ECG data of CPSC-2018, with good antinoise ability, which makes the method have a more competitive performance than other deep learning algorithms. The source code is publicly available at https://github.com/HeartbeatAI/Deform-CNN . [ABSTRACT FROM AUTHOR]