Convolutional Neural Networks for patient-specific ECG classification

An Intersubject Brain-Computer Interface Based on Domain-Adversarial Training of Convolutional Neural Network

OBJECTIVE

Attention decoding plays a vital role in daily life, where electroencephalography (EEG) has been widely involved. However, training a universally effective model for everyone is impractical due to substantial interindividual variability in EEG signals. To tackle the above challenge, we propose an end-to-end brain-computer interface (BCI) framework, including temporal and spatial one-dimensional (1D) convolutional neural network and domain-adversarial training strategy, namely DA-TSnet.

METHOD

Specifically, DA-TSnet extracts temporal and spatial features of EEG, while it is jointly supervised by task loss and domain loss. During training, DA-TSnet aims to maximize the domain loss while simultaneously minimizing the task loss. We conduct an offline analysis, simulate online experiments on a self-collected dataset of 85 subjects, and real online experiments on 22 subjects.

MAIN RESULTS

DA-TSnet achieves a leave-one-subject-out (LOSO) cross-validation (CV) classification accuracy of 89.40% ± 9.96%, outperforming several state-of-the-art attention EEG decoding methods. In simulated online experiments, DA-TSnet achieves an outstanding accuracy of 88.07% ± 11.22%. In real online experiments, it achieves an average accuracy surpassing 86%.

SIGNIFICANCE

An end-to-end network framework does not rely on elaborate preprocessing and feature extraction steps, which saves time and human workload. Moreover, our framework utilizes domain-adversarial training neural network (DANN) to tackle the challenge posed by the high interindividual variability in EEG signals, which has significant reference value for handling other EEG signal decoding issues. Last, the performance of the DA-TSnet framework in offline and online experiments underscores its potential to facilitate more reliable applications.

Review on spiking neural network-based ECG classification methods for low-power environments

This paper reviews arrhythmia classification studies using electrocardiogram (ECG) signals. Research on automatically diagnosing arrhythmia in daily life has been actively underway for early detection and treatment of heart disease. Development of automatic arrhythmia classification using ECG signal began based on handcrafted morphological feature extraction and machine learning-based classification methods. As deep neural networks (DNN) show excellent performance in the signal processing field, studies using various types of DNN are also being conducted in ECG classification. However, these DNN-based studies have extremely high computational complexity, making it challenging to perform real-time classification, and are unsuitable for low-power environments such as wearable devices due to high power consumption. Currently, research based on spiking neural network (SNN), which mimics the low-power operation of the human nervous system, is attracting attention as a method that can dramatically reduce complexity and power consumption. The classification accuracy of the SNN-based ECG classification studies is close to that of the DNN-based studies. When combined with neuromorphic hardware, it shows ultra-low-power performance, suggesting the possibility of use in lightweight devices. In this paper, the SNN-based ECG classification studies for low-power environments are mainly reviewed, and prior to this, conventional and DNN-based ECG classification studies are also reviewed. We hope that this review will be helpful to researchers and engineers interested in the field of ECG classification.

Cardiac Multi-Frequency Vibration Signal Sensor Module and Feature Extraction Method Based on Vibration Modeling

Cardiovascular diseases pose a long-term risk to human health. This study focuses on the rich-spectrum mechanical vibrations generated during cardiac activity. By combining Fourier series theory, we propose a multi-frequency vibration model for the heart, decomposing cardiac vibration into frequency bands and establishing a systematic interpretation for detecting multi-frequency cardiac vibrations. Based on this, we develop a small multi-frequency vibration sensor module based on flexible polyvinylidene fluoride (PVDF) films, which is capable of synchronously collecting ultra-low-frequency seismocardiography (ULF-SCG), seismocardiography (SCG), and phonocardiography (PCG) signals with high sensitivity. Comparative experiments validate the sensor's performance and we further develop an algorithm framework for feature extraction based on 1D-CNN models, achieving continuous recognition of multiple vibration features. Testing shows that the recognition coefficient of determination (R2), mean absolute error (MAE), and root mean square error (RMSE) of the 8 features are 0.95, 2.18 ms, and 4.89 ms, respectively, with an average prediction speed of 60.18 us/point, meeting the re-quirements for online monitoring while ensuring accuracy in extracting multiple feature points. Finally, integrating the vibration model, sensor, and feature extraction algorithm, we propose a dynamic monitoring system for multi-frequency cardiac vibration, which can be applied to portable monitoring devices for daily dynamic cardiac monitoring, providing a new approach for the early diagnosis and prevention of cardiovascular diseases.

Efficient automated error detection in medical data using deep-learning and label-clustering

Medical datasets inherently contain errors from subjective or inaccurate test results, or from confounding biological complexities. It is difficult for medical experts to detect these elusive errors manually, due to lack of contextual information, limiting data privacy regulations, and the sheer scale of data to be reviewed. Current methods for training robust artificial intelligence (AI) models on data containing mislabeled examples generally fall into one of several categories-attempting to improve the robustness of the model architecture, the regularization techniques used, the loss function used during training, or selecting a subset of data that contains cleaner labels. This last category requires the ability to efficiently detect errors either prior to or during training, either relabeling them or removing them completely. More recent progress in error detection has focused on using multi-network learning to minimize deleterious effects of errors on training, however, using many neural networks to reach a consensus on which data should be removed can be computationally intensive and inefficient. In this work, a deep-learning based algorithm was used in conjunction with a label-clustering approach to automate error detection. For dataset with synthetic label flips added, these errors were identified with an accuracy of up to 85%, while requiring up to 93% less computing resources to complete compared to a previous model consensus approach developed previously. The resulting trained AI models exhibited greater training stability and up to a 45% improvement in accuracy, from 69 to over 99% compared to the consensus approach, at least 10% improvement on using noise-robust loss functions in a binary classification problem, and a 51% improvement for multi-class classification. These results indicate that practical, automated a priori detection of errors in medical data is possible, without human oversight.

Robust Peak Detection for Holter ECGs by Self-Organized Operational Neural Networks

Although numerous R-peak detectors have been proposed in the literature, their robustness and performance levels may significantly deteriorate in low-quality and noisy signals acquired from mobile electrocardiogram (ECG) sensors, such as Holter monitors. Recently, this issue has been addressed by deep 1-D convolutional neural networks (CNNs) that have achieved state-of-the-art performance levels in Holter monitors; however, they pose a high complexity level that requires special parallelized hardware setup for real-time processing. On the other hand, their performance deteriorates when a compact network configuration is used instead. This is an expected outcome as recent studies have demonstrated that the learning performance of CNNs is limited due to their strictly homogenous configuration with the sole linear neuron model. This has been addressed by operational neural networks (ONNs) with their heterogenous network configuration encapsulating neurons with various nonlinear operators. In this study, to further boost the peak detection performance along with an elegant computational efficiency, we propose 1-D Self-Organized ONNs (Self-ONNs) with generative neurons. The most crucial advantage of 1-D Self-ONNs over the ONNs is their self-organization capability that voids the need to search for the best operator set per neuron since each generative neuron has the ability to create the optimal operator during training. The experimental results over the China Physiological Signal Challenge-2020 (CPSC) dataset with more than one million ECG beats show that the proposed 1-D Self-ONNs can significantly surpass the state-of-the-art deep CNN with less computational complexity. Results demonstrate that the proposed solution achieves a 99.10% F1-score, 99.79% sensitivity, and 98.42% positive predictivity in the CPSC dataset, which is the best R-peak detection performance ever achieved.

Deep Learning for Structural Health Monitoring: Data, Algorithms, Applications, Challenges, and Trends

Environmental effects may lead to cracking, stiffness loss, brace damage, and other damages in bridges, frame structures, buildings, etc. Structural Health Monitoring (SHM) technology could prevent catastrophic events by detecting damage early. In recent years, Deep Learning (DL) has developed rapidly and has been applied to SHM to detect, localize, and evaluate diverse damages through efficient feature extraction. This paper analyzes 337 articles through a systematic literature review to investigate the application of DL for SHM in the operation and maintenance phase of facilities from three perspectives: data, DL algorithms, and applications. Firstly, the data types in SHM and the corresponding collection methods are summarized and analyzed. The most common data types are vibration signals and images, accounting for 80% of the literature studied. Secondly, the popular DL algorithm types and application areas are reviewed, of which CNN accounts for 60%. Then, this article carefully analyzes the specific functions of DL application for SHM based on the facility's characteristics. The most scrutinized study focused on cracks, accounting for 30 percent of research papers. Finally, challenges and trends in applying DL for SHM are discussed. Among the trends, the Structural Health Monitoring Digital Twin (SHMDT) model framework is suggested in response to the trend of strong coupling between SHM technology and Digital Twin (DT), which can advance the digitalization, visualization, and intelligent management of SHM.

Design and Evaluation of Deep Learning Models for Continuous Acute Pain Detection Based on Phasic Electrodermal Activity

The current method for assessing pain in clinical practice is subjective and relies on self-reported scales. An objective and accurate method of pain assessment is needed for physicians to prescribe the proper medication dosage, which could reduce addiction to opioids. Hence, many works have used electrodermal activity (EDA) as a suitable signal for detecting pain. Previous studies have used machine learning and deep learning to detect pain responses, but none have used a sequence-to-sequence deep learning approach to continuously detect acute pain from EDA signals, as well as accurate detection of pain onset. In this study, we evaluated deep learning models including 1-dimensional convolutional neural networks (1D-CNN), long short-term memory networks (LSTM), and three hybrid CNN-LSTM architectures for continuous pain detection using phasic EDA features. We used a database consisting of 36 healthy volunteers who underwent pain stimuli induced by a thermal grill. We extracted the phasic component, phasic drivers, and time-frequency spectrum of the phasic EDA (TFS-phEDA), which was found to be the most discerning physiomarker. The best model was a parallel hybrid architecture of a temporal convolutional neural network and a stacked bi-directional and uni-directional LSTM, which obtained a F1-score of 77.8% and was able to correctly detect pain in 15-second signals. The model was evaluated using 37 independent subjects from the BioVid Heat Pain Database and outperformed other approaches in recognizing higher pain levels compared to baseline with an accuracy of 91.5%. The results show the feasibility of continuous pain detection using deep learning and EDA.

Identification and Classification of Human Body Exercises on Smart Textile Bands by Combining Decision Tree and Convolutional Neural Networks

In recent years, human activity recognition (HAR) has gained significant interest from researchers in the sports and fitness industries. In this study, the authors have proposed a cascaded method including two classifying stages to classify fitness exercises, utilizing a decision tree as the first stage and a one-dimension convolutional neural network as the second stage. The data acquisition was carried out by five participants performing exercises while wearing an inertial measurement unit sensor attached to a wristband on their wrists. However, only data acquired along the z-axis of the IMU accelerator was used as input to train and test the proposed model, to simplify the model and optimize the training time while still achieving good performance. To examine the efficiency of the proposed method, the authors compared the performance of the cascaded model and the conventional 1D-CNN model. The obtained results showed an overall improvement in the accuracy of exercise classification by the proposed model, which was approximately 92%, compared to 82.4% for the 1D-CNN model. In addition, the authors suggested and evaluated two methods to optimize the clustering outcome of the first stage in the cascaded model. This research demonstrates that the proposed model, with advantages in terms of training time and computational cost, is able to classify fitness workouts with high performance. Therefore, with further development, it can be applied in various real-time HAR applications.

Detection of Cardiovascular Disease from Clinical Parameters Using a One-Dimensional Convolutional Neural Network

Heart disease is a significant public health problem, and early detection is crucial for effective treatment and management. Conventional and noninvasive techniques are cumbersome, time-consuming, inconvenient, expensive, and unsuitable for frequent measurement or diagnosis. With the advance of artificial intelligence (AI), new invasive techniques emerging in research are detecting heart conditions using machine learning (ML) and deep learning (DL). Machine learning models have been used with the publicly available dataset from the internet about heart health; in contrast, deep learning techniques have recently been applied to analyze electrocardiograms (ECG) or similar vital data to detect heart diseases. Significant limitations of these datasets are their small size regarding the number of patients and features and the fact that many are imbalanced datasets. Furthermore, the trained models must be more reliable and accurate in medical settings. This study proposes a hybrid one-dimensional convolutional neural network (1D CNN), which uses a large dataset accumulated from online survey data and selected features using feature selection algorithms. The 1D CNN proved to show better accuracy compared to contemporary machine learning algorithms and artificial neural networks. The non-coronary heart disease (no-CHD) and CHD validation data showed an accuracy of 80.1% and 76.9%, respectively. The model was compared with an artificial neural network, random forest, AdaBoost, and a support vector machine. Overall, 1D CNN proved to show better performance in terms of accuracy, false negative rates, and false positive rates. Similar strategies were applied for four more heart conditions, and the analysis proved that using the hybrid 1D CNN produced better accuracy.

Hardware-Efficient 1D CNN for Patient-Specific Early Seizure Detection

Closed-loop brain-implantable neuromodulation devices are a new treatment option for patients with refractory epilepsy. Seizure detection algorithms implemented on such devices are subject to strict power and area constraints. Deep learning methods, though very powerful, tend to have high computational complexity and thus are typically impractical for resource-constrained neuromodulation devices. In this paper, we propose a compact and hardware-efficient one-dimensional convolutional neural network (1D CNN) structure for patient-specific early seizure detection. Feature extraction techniques and a novel initialization method based on the forward-chaining training and testing scheme are used to improve model performance. Our compact model achieves similar accuracy to that of support vector machines, the state-of-the-art method for seizure detection, while consuming over 20x less power.

Classification of Breathing Signals According to Human Motions by Combining 1D Convolutional Neural Network and Embroidered Textile Sensor

Research on healthcare and body monitoring has increased in recent years, with respiratory data being one of the most important factors. Respiratory measurements can help prevent diseases and recognize movements. Therefore, in this study, we measured respiratory data using a capacitance-based sensor garment with conductive electrodes. To determine the most stable measurement frequency, we conducted experiments using a porous Eco-flex and selected 45 kHz as the most stable frequency. Next, we trained a 1D convolutional neural network (CNN) model, which is a type of deep learning model, to classify the respiratory data according to four movements (standing, walking, fast walking, and running) using one input. The final test accuracy for classification was >95%. Therefore, the sensor garment developed in this study can measure respiratory data for four movements and classify them using deep learning, making it a versatile wearable in the form of a textile. We expect that this method will advance in various healthcare fields.

Real-Time Liquid Rate and Water Cut Prediction From the Electrical Submersible Pump Sensors Data Using Machine-Learning Algorithms

This study presents a novel data-driven approach for calculating multiphase flow rates in electrical submersible pumped wells. Traditional methods for estimating flow rates at test separators fail to identify production trends and require additional costs for maintenance. In response, virtual flow metering (VFM) has emerged as an attractive research area in the oil and gas industry. This study introduces a robust workflow utilizing symbolic regression, extreme gradient boosted trees, and a deep learning model that includes a pipeline of convolutional neural network (CNN) layers and long short-term memory algorithm (LSTM) layers to predict liquid rate and water cut in real time based on pump sensors' data. The novelty of this approach lies in offering a cost-effective and accurate alternative to the usage of multiphase physical flow meters and production testing. Additionally, the study provides insights into the potential of data-driven methods for VFM in electrical submersible pumped wells, highlighting the effectiveness of the proposed approach. Overall, this study contributes to the field by introducing a new, data-driven method for accurately predicting multiphase flow rates in real time, thereby providing a valuable tool for monitoring and optimizing production processes in the oil and gas industry.

The prediction of cardiac abnormality and enhancement in minority class accuracy from imbalanced ECG signals using modified deep neural network models

Cardiovascular disease (CVD) is the most fatal disease in the world, so its accurate and automated detection in the early stages will certainly support the medical expert in timely diagnosis and treatment, which can save many lives. Many types of research have been carried out in this regard, but due to the problem of data imbalance in the medical and health care sector, it may not provide the desired results in all aspects. To overcome this problem, a sequential ensemble technique has been proposed that detects 6 types of cardiac arrhythmias on large ECG imbalanced datasets, and the data imbalanced issue of the ECG dataset has been addressed by using a hybrid data resampling technique called "Synthetically Minority Oversampling Technique and Tomek Link (SMOTE + Tomek)". The sequential ensemble technique employs two distinct deep learning models: Convolutional Neural Network (CNN) and a hybrid model, CNN with Long Short-Term Memory Network (CNN-LSTM). The two standard datasets "MIT-BIH arrhythmias database" (MITDB) and "PTB diagnostic database" (PTBDB) are combined and extracted 23, 998 ECG beats for the model validation. In this work, the three models CNN, CNN-LSTM, and ensemble approach were tested on four kinds of ECG datasets: the original data (imbalanced), the data sampled using a random oversampled technique, data sampled using SMOTE, and the dataset resampled using SMOTE + Tomek algorithm. The overall highest accuracy was obtained of 99.02% on the SMOTE + Tomek sampled dataset by ensemble technique and the minority class accuracy result (Recall) is improved by 20% as compared to the imbalanced data.

Low Complexity Binarized 2D-CNN Classifier for Wearable Edge AI Devices

Wearable Artificial Intelligence-of-Things (AIoT) devices exhibit the need to be resource and energy-efficient. In this paper, we introduced a quantized multilayer perceptron (qMLP) for converting ECG signals to binary image, which can be combined with binary convolutional neural network (bCNN) for classification. We deploy our model into a low-power and low-resource field programmable gate array (FPGA) fabric. The model requires 5.8× lesser multiply and accumulate (MAC) operations than known wearable CNN models. Our model also achieves a classification accuracy of 98.5%, sensitivity of 85.4%, specificity of 99.5%, precision of 93.3%, and F1-score of 89.2%, along with dynamic power dissipation of 34.9 μW.

Discriminating the occurrence of inundation in tsunami early warning with one-dimensional convolutional neural networks

Tsunamis are natural phenomena that, although occasional, can have large impacts on coastal environments and settlements, especially in terms of loss of life. An accurate, detailed and timely assessment of the hazard is essential as input for mitigation strategies both in the long term and during emergencies. This goal is compounded by the high computational cost of simulating an adequate number of scenarios to make robust assessments. To reduce this handicap, alternative methods could be used. Here, an enhanced method for estimating tsunami time series using a one-dimensional convolutional neural network model (1D CNN) is considered. While the use of deep learning for this problem is not new, most of existing research has focused on assessing the capability of a network to reproduce inundation metrics extrema. However, for the context of Tsunami Early Warning, it is equally relevant to assess whether the networks can accurately predict whether inundation would occur or not, and its time series if it does. Hence, a set of 6776 scenarios with magnitudes in the range [Formula: see text] 8.0-9.2 were used to design several 1D CNN models at two bays that have different hydrodynamic behavior, that would use as input inexpensive low-resolution numerical modeling of tsunami propagation to predict inundation time series at pinpoint locations. In addition, different configuration parameters were also analyzed to outline a methodology for model testing and design, that could be applied elsewhere. The results show that the network models are capable of reproducing inundation time series well, either for small or large flow depths, but also when no inundation was forecast, with minimal instances of false alarms or missed alarms. To further assess the performance, the model was tested with two past tsunamis and compared with actual inundation metrics. The results obtained are promising, and the proposed model could become a reliable alternative for the calculation of tsunami intensity measures in a faster than real time manner. This could complement existing early warning system, by means of an approximate and fast procedure that could allow simulating a larger number of scenarios within the always restricting time frame of tsunami emergencies.

Deep Learning and Hyperspectral Images Based Tomato Soluble Solids Content and Firmness Estimation

Cherry tomato (Solanum lycopersicum) is popular with consumers over the world due to its special flavor. Soluble solids content (SSC) and firmness are two key metrics for evaluating the product qualities. In this work, we develop non-destructive testing techniques for SSC and fruit firmness based on hyperspectral images and the corresponding deep learning regression model. Hyperspectral reflectance images of over 200 tomato fruits are derived with the spectrum ranging from 400 to 1,000 nm. The acquired hyperspectral images are corrected and the spectral information are extracted. A novel one-dimensional (1D) convolutional ResNet (Con1dResNet) based regression model is proposed and compared with the state of art techniques. Experimental results show that, with a relatively large number of samples our technique is 26.4% better than state of art technique for SSC and 33.7% for firmness. The results of this study indicate the application potential of hyperspectral imaging technique in the SSC and firmness detection, which provides a new option for non-destructive testing of cherry tomato fruit quality in the future.

Local Discrimination Based on Piezoelectric Sensing in Robots Composed of Soft Matter with Different Physical Properties

The coronavirus epidemic has attracted significant attention to the applications of pet robots which can be used to treat and entertain people in their homes. However, pet robots are fabricated using hard materials and it is difficult for them to communicate with people through contact. Soft robots are expected to realize communication through contact similar to that of actual pets. Soft robots provide people with a sense of healing and security owing to their softness and can extract rich information through external stimuli by applying a machine learning framework called physical-reservoir computing. It is crucial to determine the differences between the physical properties of soft materials that affect the information extracted from a soft body to develop an intelligent soft robot. In this study, two owl-shaped soft robots with different softnesses were developed to analyze the characteristics of the signal data obtained via piezoelectric film sensors embedded in models with different physical properties. An accuracy of 94.2% and 95.9% was obtained for touched part classification using 1D CNN and logistic regression models, respectively. Additionally, the relationship between the softness of material and classification performance was investigated by comparing the distribution of part classification accuracy for different hyper-parameters of two owl models.

An Energy Efficient ECG Ventricular Ectopic Beat Classifier Using Binarized CNN for Edge AI Devices

Wearable Artificial Intelligence-of-Things (AIoT) requires edge devices to be resource and energy-efficient. In this paper, we design and implement an efficient binary convolutional neural network (bCNN) algorithm utilizing function-merging and block-reuse techniques to classify between Ventricular and non-Ventricular Ectopic Beat images. We deploy our model into a low-resource low-power field programmable gate array (FPGA) fabric. Our model achieves a classification accuracy of 97.3%, sensitivity of 91.3%, specificity of 98.1%, precision of 86.7%, and F1-score of 88.9%, along with dynamic power dissipation of only 10.5-μW.

Real-Time Patient-Specific ECG Classification by 1D Self-Operational Neural Networks

Despite the proliferation of numerous deep learning methods proposed for generic ECG classification and arrhythmia detection, compact systems with the real-time ability and high accuracy for classifying patient-specific ECG are still few. Particularly, the scarcity of patient-specific data poses an ultimate challenge to any classifier. Recently, compact 1D Convolutional Neural Networks (CNNs) have achieved the state-of-the-art performance level for the accurate classification of ventricular and supraventricular ectopic beats. However, several studies have demonstrated the fact that the learning performance of the conventional CNNs is limited because they are homogenous networks with a basic (linear) neuron model. In order to address this deficiency and further boost the patient-specific ECG classification performance, in this study, we propose 1D Self-organized Operational Neural Networks (1D Self-ONNs). Due to its self-organization capability, Self-ONNs have the utmost advantage and superiority over conventional ONNs where the prior operator search within the operator set library to find the best possible set of operators is entirely avoided. As the first study where 1D Self-ONNs are ever proposed for a classification task, our results over the MIT-BIH arrhythmia benchmark database demonstrate that 1D Self-ONNs can surpass 1D CNNs with a significant margin while having a similar computational complexity. Under AAMI recommendations and with minimal common training data used, over the entire MIT-BIH dataset 1D Self-ONNs have achieved 98% and 99.04% average accuracies, 76.6% and 93.7% average F1 scores on supra-ventricular and ventricular ectopic beat (VEB) classifications, respectively, which is the highest performance level ever reported.

On the Transition of Social Interaction from In-Person to Online: Predicting Changes in Social Media Usage of College Students during the COVID-19 Pandemic based on Pre-COVID-19 On-Campus Colocation

Pandemics significantly impact human daily life. People throughout the world adhere to safety protocols (e.g., social distancing and self-quarantining). As a result, they willingly keep distance from workplace, friends and even family. In such circumstances, in-person social interactions may be substituted with virtual ones via online channels, such as, Instagram and Snapchat. To get insights into this phenomenon, we study a group of undergraduate students before and after the start of COVID-19 pandemic. Specifically, we track N=102 undergraduate students on a small college campus prior to the pandemic using mobile sensing from phones and assign semantic labels to each location they visit on campus where they study, socialize and live. By leveraging their colocation network at these various semantically labeled places on campus, we find that colocations at certain places that possibly proxy higher in-person social interactions (e.g., dormitories, gyms and Greek houses) show significant predictive capability in identifying the individuals' change in social media usage during the pandemic period. We show that we can predict student's change in social media usage during COVID-19 with an F1 score of 0.73 purely from the in-person colocation data generated prior to the pandemic.

A Novel Interference Avoidance Based on a Distributed Deep Learning Model for 5G-Enabled IoT

The co-existence of fifth-generation (5G) and Internet-of-Things (IoT) has become inevitable in many applications since 5G networks have created steadier connections and operate more reliably, which is extremely important for IoT communication. During transmission, IoT devices (IoTDs) communicate with IoT Gateway (IoTG), whereas in 5G networks, cellular users equipment (CUE) may communicate with any destination (D) whether it is a base station (BS) or other CUE, which is known as device-to-device (D2D) communication. One of the challenges that face 5G and IoT is interference. Interference may exist at BSs, CUE receivers, and IoTGs due to the sharing of the same spectrum. This paper proposes an interference avoidance distributed deep learning model for IoT and device to any destination communication by learning from data generated by the Lagrange optimization technique to predict the optimum IoTD-D, CUE-IoTG, BS-IoTD and IoTG-CUE distances for uplink and downlink data communication, thus achieving higher overall system throughput and energy efficiency. The proposed model was compared to state-of-the-art regression benchmarks, which provided a huge improvement in terms of mean absolute error and root mean squared error. Both analytical and deep learning models reached the optimal throughput and energy efficiency while suppressing interference to any destination and IoTG.

Recognition of Fine-Grained Walking Patterns Using a Smartwatch with Deep Attentive Neural Networks

Generally, people do various things while walking. For example, people frequently walk while looking at their smartphones. Sometimes we walk differently than usual; for example, when walking on ice or snow, we tend to waddle. Understanding walking patterns could provide users with contextual information tailored to the current situation. To formulate this as a machine-learning problem, we defined 18 different everyday walking styles. Noting that walking strategies significantly affect the spatiotemporal features of hand motions, e.g., the speed and intensity of the swinging arm, we propose a smartwatch-based wearable system that can recognize these predefined walking styles. We developed a wearable system, suitable for use with a commercial smartwatch, that can capture hand motions in the form of multivariate timeseries (MTS) signals. Then, we employed a set of machine learning algorithms, including feature-based and recent deep learning algorithms, to learn the MTS data in a supervised fashion. Experimental results demonstrated that, with recent deep learning algorithms, the proposed approach successfully recognized a variety of walking patterns, using the smartwatch measurements. We analyzed the results with recent attention-based recurrent neural networks to understand the relative contributions of the MTS signals in the classification process.

A Lightweight CNN Model for Detecting Respiratory Diseases From Lung Auscultation Sounds Using EMD-CWT-Based Hybrid Scalogram

Listening to lung sounds through auscultation is vital in examining the respiratory system for abnormalities. Automated analysis of lung auscultation sounds can be beneficial to the health systems in low-resource settings where there is a lack of skilled physicians. In this work, we propose a lightweight convolutional neural network (CNN) architecture to classify respiratory diseases from individual breath cycles using hybrid scalogram-based features of lung sounds. The proposed feature-set utilizes the empirical mode decomposition (EMD) and the continuous wavelet transform (CWT). The performance of the proposed scheme is studied using a patient independent train-validation-test set from the publicly available ICBHI 2017 lung sound dataset. Employing the proposed framework, weighted accuracy scores of 98.92% for three-class chronic classification and 98.70% for six-class pathological classification are achieved, which outperform well-known and much larger VGG16 in terms of accuracy by absolute margins of 1.10% and 1.11%, respectively. The proposed CNN model also outperforms other contemporary lightweight models while being computationally comparable.

Estimation of Organizational Competitiveness by a Hybrid of One-Dimensional Convolutional Neural Networks and Self-Organizing Maps Using Physiological Signals for Emotional Analysis of Employees

The theory of modern organizations considers emotional intelligence to be the metric for tools that enable organizations to create a competitive vision. It also helps corporate leaders enthusiastically adhere to the vision and energize organizational stakeholders to accomplish the vision. In this study, the one-dimensional convolutional neural network classification model is initially employed to interpret and evaluate shifts in emotion over a period by categorizing emotional states that occur at particular moments during mutual interaction using physiological signals. The self-organizing map technique is implemented to cluster overall organizational emotions to represent organizational competitiveness. The analysis of variance test results indicates no significant difference in age and body mass index for participants exhibiting different emotions. However, a significant mean difference was observed for the blood volume pulse, galvanic skin response, skin temperature, valence, and arousal values, indicating the effectiveness of the chosen physiological sensors and their measures to analyze emotions for organizational competitiveness. We achieved 99.8% classification accuracy for emotions using the proposed technique. The study precisely identifies the emotions and locates a connection between emotional intelligence and organizational competitiveness (i.e., a positive relationship with employees augments organizational competitiveness).

Comparative analysis between convolutional neural network learned and engineered features: A case study on cardiac arrhythmia detection

Background

Atrial fibrillation (AF) is one of the most common cardiovascular problems, and its asymptomatic tendency makes AF detection challenging. Machine and deep learning methods are commonly used in AF detection.

Objective

The purpose of this study was to evaluate the information provided by convolutional neural network (CNN) and random forest (RF) machine learning models for AF classification.

Methods

We manually extracted 166 time-frequency domains and linear and nonlinear features to classify single-lead electrocardiograms (ECGs) as normal, AF, other, or noisy sinus rhythms. We selected a subset of 56 robust features using a genetic algorithm that was used in the RF model. In a separate study, a 1-dimensional, 12-layer CNN was designed on the raw ECG rhythms. Four features from the output layer and 128 features from the fully connected layer of CNN were explored independently for classification. The models were trained and internally validated on 8,528 ECGs and externally validated on a hidden dataset containing 3,658 ECGs. Next,we analyzed the correlation between engineered and CNN-learned features.

Results

An RF classifier trained with 56-engineered features resulted in an F₁ score of 0.91, 0.78, and 0.72 for normal, AF, and other rhythms, respectively. However, an ensemble of support vector machine and the CNN model resulted in an F₁ score of 0.92, 0.87, and 0.80, respectively.

Conclusion

We explored various features and machine learning models to identify AF rhythms using short (9-61 seconds) single-lead ECG recordings. Our results showed that the proposed CNN model abstracted distinctive features for AF classification.

Deep Neural Network for Respiratory Sound Classification in Wearable Devices Enabled by Patient Specific Model Tuning

The primary objective of this paper is to build classification models and strategies to identify breathing sound anomalies (wheeze, crackle) for automated diagnosis of respiratory and pulmonary diseases. In this work we propose a deep CNN-RNN model that classifies respiratory sounds based on Mel-spectrograms. We also implement a patient specific model tuning strategy that first screens respiratory patients and then builds patient specific classification models using limited patient data for reliable anomaly detection. Moreover, we devise a local log quantization strategy for model weights to reduce the memory footprint for deployment in memory constrained systems such as wearable devices. The proposed hybrid CNN-RNN model achieves a score of [Formula: see text] on four-class classification of breathing cycles for ICBHI'17 scientific challenge respiratory sound database. When the model is re-trained with patient specific data, it produces a score of [Formula: see text] for leave-one-out validation. The proposed weight quantization technique achieves ≈ 4 × reduction in total memory cost without loss of performance. The main contribution of the paper is as follows: Firstly, the proposed model is able to achieve state of the art score on the ICBHI'17 dataset. Secondly, deep learning models are shown to successfully learn domain specific knowledge when pre-trained with breathing data and produce significantly superior performance compared to generalized models. Finally, local log quantization of trained weights is shown to be able to reduce the memory requirement significantly. This type of patient-specific re-training strategy can be very useful in developing reliable long-term automated patient monitoring systems particularly in wearable healthcare solutions.

A Survey of Heart Anomaly Detection Using Ambulatory Electrocardiogram (ECG)

Cardiovascular diseases (CVDs) are the number one cause of death globally. An estimated 17.9 million people die from CVDs each year, representing 31% of all global deaths. Most cardiac patients require early detection and treatment. Therefore, many products to monitor patient's heart conditions have been introduced on the market. Most of these devices can record a patient's bio-metric signals both in resting and in exercising situations. However, reading the massive amount of raw electrocardiogram (ECG) signals from the sensors is very time-consuming. Automatic anomaly detection for the ECG signals could act as an assistant for doctors to diagnose a cardiac condition. This paper reviews the current state-of-the-art of this technology discusses the pros and cons of the devices and algorithms found in the literature and the possible research directions to develop the next generation of ambulatory monitoring systems.

Transfer Learning in ECG Classification from Human to Horse Using a Novel Parallel Neural Network Architecture

Automatic or semi-automatic analysis of the equine electrocardiogram (eECG) is currently not possible because human or small animal ECG analysis software is unreliable due to a different ECG morphology in horses resulting from a different cardiac innervation. Both filtering, beat detection to classification for eECGs are currently poorly or not described in the literature. There are also no public databases available for eECGs as is the case for human ECGs. In this paper we propose the use of wavelet transforms for both filtering and QRS detection in eECGs. In addition, we propose a novel robust deep neural network using a parallel convolutional neural network architecture for ECG beat classification. The network was trained and tested using both the MIT-BIH arrhythmia and an own made eECG dataset with 26.440 beats on 4 classes: normal, premature ventricular contraction, premature atrial contraction and noise. The network was optimized using a genetic algorithm and an accuracy of 97.7% and 92.6% was achieved for the MIT-BIH and eECG database respectively. Afterwards, transfer learning from the MIT-BIH dataset to the eECG database was applied after which the average accuracy, recall, positive predictive value and F1 score of the network increased with an accuracy of 97.1%.

Automatic detection of arrhythmia from imbalanced ECG database using CNN model with SMOTE

Timely prediction of cardiovascular diseases with the help of a computer-aided diagnosis system minimizes the mortality rate of cardiac disease patients. Cardiac arrhythmia detection is one of the most challenging tasks, because the variations of electrocardiogram(ECG) signal are very small, which cannot be detected by human eyes. In this study, an 11-layer deep convolutional neural network model is proposed for classification of the MIT-BIH arrhythmia database into five classes according to the ANSI-AAMI standards. In this CNN model, we designed a complete end-to-end structure of the classification method and applied without the denoising process of the database. The major advantage of the new methodology proposed is that the number of classifications will reduce and also the need to detect, and segment the QRS complexes, obviated. This MIT-BIH database has been artificially oversampled to handle the minority classes, class imbalance problem using SMOTE technique. This new CNN model was trained on the augmented ECG database and tested on the real dataset. The experimental results portray that the developed CNN model has better performance in terms of precision, recall, F-score, and overall accuracy as compared to the work mentioned in the literatures. These results also indicate that the best performance accuracy of 98.30% is obtained in the 70:30 train-test data set.

A deep neural network for 12-lead electrocardiogram interpretation outperforms a conventional algorithm, and its physician overread, in the diagnosis of atrial fibrillation

Background

Automated electrocardiogram (ECG) interpretations may be erroneous, and lead to erroneous overreads, including for atrial fibrillation (AF). We compared the accuracy of the first version of a new deep neural network 12-Lead ECG algorithm (Cardiologs®) to the conventional Veritas algorithm in interpretation of AF.

Methods

24,123 consecutive 12-lead ECGs recorded over 6 months were interpreted by 1) the Veritas® algorithm, 2) physicians who overread Veritas® (Veritas® + physician), and 3) Cardiologs® algorithm. We randomly selected 500 out of 858 ECGs with a diagnosis of AF according to either algorithm, then compared the algorithms' interpretations, and Veritas® + physician, with expert interpretation. To assess sensitivity for AF, we analyzed a separate database of 1473 randomly selected ECGs interpreted by both algorithms and by blinded experts.

Results

Among the 500 ECGs selected, 399 had a final classification of AF; 101 (20.2%) had ≥1 false positive automated interpretation. Accuracy of Cardiologs® (91.2%; CI: 82.4–94.4) was higher than Veritas® (80.2%; CI: 76.5–83.5) (p < 0.0001), and equal to Veritas® + physician (90.0%, CI:87.1–92.3) (p = 0.12). When Veritas® was incorrect, accuracy of Veritas® + physician was only 62% (CI 52–71); among those ECGs, Cardiologs® accuracy was 90% (CI: 82–94; p < 0.0001). The second database had 39 AF cases; sensitivity was 92% vs. 87% (p = 0.46) and specificity was 99.5% vs. 98.7% (p = 0.03) for Cardiologs® and Veritas® respectively.

Conclusion

Cardiologs® 12-lead ECG algorithm improves the interpretation of atrial fibrillation.

A novel method to monitor human stress states using ultra-short-term ECG spectral feature

Electrocardiogram (ECG) signal represents autonomous nervous system responses to human emotional states. This research demonstrates that the spectral ECG features within ultra-short-term window duration (10-sec) could be utilized to monitor human emotional states. Experiments were conducted with five different stress protocols including mental and physical tasks. Experimental results showed feasible classification performance of ECG spectral features compared to that of HRV parameters. The averaged classification accuracy across 13 subjects and all stress protocols was 81.16% using Naïve Bayes algorithm. In addition, the results showed stress responses in mental arithmetic tasks was the most separable from those in resting states (87.31%) compared to the other stress situations.