Search

This paper describes the validation of a deep learning model for Internet of Things (IoT) based health care applications. As such, the deep learning model was created to detect episodes of Atrial Fibrillation (AF) using Heart Rate (HR) signals. The initial Long Short-Term Memory (LSTM) model was developed using 20 data sets, from distinct subjects, obtained from the AFDB database on PhysioNet. This model achieved an AF detection accuracy of 98.51% with ten fold cross validation. In this study, we validated the initial results by testing the developed deep learning model with unknown data. To be specific, we fed the data from 82 subjects to the deep learning system and compared the classification results with the diagnosis results indicated by human practitioners. The validation results show 94% accuracy with an area under the Receiver Operating Characteristic (ROC) curve of 96.58. These results indicate that the LSTM model is able to extract the feature maps from the unknown data and hence detect the AF periods accurately. With this blindfold validation testing we violated a well known design rule for learning systems which states that more data should be used for training than for testing. By doing so, we have established that our deep learning system is fit for practical deployment, because in a practical situation the diagnosis support system must apply the knowledge, extracted from a limited training data set, to a HR trace from a patient.

Credit scoring (CS) is an effective and crucial approach used for risk management in banks and other financial institutions. It provides appropriate guidance on granting loans and reduces risks in the financial area. Hence, companies and banks are trying to use novel automated solutions to deal with CS challenge to protect their own finances and customers. Nowadays, different machine learning (ML) and data mining (DM) algorithms have been used to improve various aspects of CS prediction. In this paper, we introduce a novel methodology, named Deep Genetic Hierarchical Network of Learners (DGHNL). The proposed methodology comprises different types of learners, including Support Vector Machines (SVM), k-Nearest Neighbors (kNN), Probabilistic Neural Networks (PNN), and fuzzy systems. The Statlog German (1000 instances) credit approval dataset available in the UCI machine learning repository is used to test the effectiveness of our model in the CS domain. Our DGHNL model encompasses five kinds of learners, two kinds of data normalization procedures, two extraction of features methods, three kinds of kernel functions, and three kinds of parameter optimizations. Furthermore, the model applies deep learning, ensemble learning, supervised training, layered learning, genetic selection of features (attributes), genetic optimization of learners parameters, and novel genetic layered training (selection of learners) approaches used along with the cross-validation (CV) training-testing method (stratified 10-fold). The novelty of our approach relies on a proper flow and fusion of information (DGHNL structure and its optimization). We show that the proposed DGHNL model with a 29-layer structure is capable to achieve the prediction accuracy of 94.60% (54 errors per 1000 classifications) for the Statlog German credit approval data. It is the best prediction performance for this well-known credit scoring dataset, compared to the existing work in the field.

The thyroid is the chief hormonal gland that controls the growth, metabolism, and maturation of the body. However, the function of the thyroid gland could be disrupted if it produces too much or too little hormones. Furthermore, there could be abnormal growth in thyroid cell tissue, leading to the formation of a benign or malignant thyroid lesion. Ultrasound is a typical non-invasive diagnosis approach to check for cancerous thyroid lesions. However, the visual interpretation of the ultrasound thyroid images is challenging and time-consuming. Hence, a feature engineering model is proposed to overcome these challenges. We propose to transform image pixel intensity values into high dimensional structured data set before fitting a Regression analysis framework to estimate kernel parameters for an image filter model. We then adopt a Bayesian network inference to estimate a subset for the textural features with a significant conditional dependency in the classification of thyroid lesions. The analysis of the proposed feature engineering model showed that the classification performance had an overall significant improvement over other image filter models. We achieve 96.00% classification accuracy with a sensitivity and specificity of 99.64% and 90.23% respectively for a filter size of 13 × 13. The analysis of results indicate that the diagnosis of ultrasound images thyroid nodules is significantly boosts by adaptively learning filter parameters for feature engineering model.

In the presence of gestational diabetes mellitus (GDM), the fetus is exposed to a hyperinsulinemia environment. This environment can cause a wide range of metabolic and fetal cardiac structural alterations. Fetal myocardial hypertrophy predominantly affecting the interventricular septum possesses a morphology of disarray similar to hypertrophic cardiomyopathy, and may be present in some GDM neonates after birth. Myocardial thickness may increase in GDM fetuses independent of glycemic control status and fetal weight. Fetal echocardiography performed on fetuses with GDM helps in assessing cardiac structure and function, and to diagnose myocardial hypertrophy. There are few studies in the literature which have established evidence for morphologic variation associated with cardiac hypertrophy among fetuses of GDM mothers. In this study, fetal ultrasound images of normal, pregestational diabetes mellitus (preGDM) and GDM mothers were used to develop a computer aided diagnostic (CAD) tool. We proposed a new method called local preserving class separation (LPCS) framework to preserve the geometrical configuration of normal and preGDM/GDM subjects. The generated shearlet based texture features under LPCS framework showed promising results compared with deep learning algorithms. The proposed method achieved a maximum accuracy of 98.15% using a support vector machine (SVM) classifier. Hence, this paradigm can be helpful to physicians in detecting fetal myocardial hypertrophy in preGDM/GDM mothers.

Myocardial infarction (MI) accounts for a high number of deaths globally. In acute MI, accurate electrocardiography (ECG) is important for timely diagnosis and intervention in the emergency setting. Machine learning is increasingly being explored for automated computer-aided ECG diagnosis of cardiovascular diseases. In this study, we have developed DenseNet and CNN models for the classification of healthy subjects and patients with ten classes of MI based on the location of myocardial involvement. ECG signals from the Physikalisch-Technische Bundesanstalt database were pre-processed, and the ECG beats were extracted using an R peak detection algorithm. The beats were then fed to the two models separately. While both models attained high classification accuracies (more than 95%), DenseNet is the preferred model for the classification task due to its low computational complexity and higher classification accuracy than the CNN model due to feature reusability. An enhanced class activation mapping (CAM) technique called Grad-CAM was subsequently applied to the outputs of both models to enable visualization of the specific ECG leads and portions of ECG waves that were most influential for the predictive decisions made by the models for the 11 classes. It was observed that Lead V4 was the most activated lead in both the DenseNet and CNN models. Furthermore, this study has also established the different leads and parts of the signal that get activated for each class. This is the first study to report features that influenced the classification decisions of deep models for multiclass classification of MI and healthy ECGs. Hence this study is crucial and contributes significantly to the medical field as with some level of visible explainability of the inner workings of the models, the developed DenseNet and CNN models may garner needed clinical acceptance and have the potential to be implemented for ECG triage of MI diagnosis in hospitals and remote out-of-hospital settings.

Mask usage is one of the most important precautions to limit the spread of COVID-19. Therefore, hygiene rules enforce the correct use of face coverings. Automated mask usage classification might be used to improve compliance monitoring. This study deals with the problem of inappropriate mask use. To address that problem, 2075 face mask usage images were collected. The individual images were labeled as either mask, no masked, or improper mask. Based on these labels, the following three cases were created: Case 1: mask versus no mask versus improper mask, Case 2: mask versus no mask + improper mask, and Case 3: mask versus no mask. This data was used to train and test a hybrid deep feature-based masked face classification model. The presented method comprises of three primary stages: (i) pre-trained ResNet101 and DenseNet201 were used as feature generators; each of these generators extracted 1000 features from an image; (ii) the most discriminative features were selected using an improved RelieF selector; and (iii) the chosen features were used to train and test a support vector machine classifier. That resulting model attained 95.95%, 97.49%, and 100.0% classification accuracy rates on Case 1, Case 2, and Case 3, respectively. Having achieved these high accuracy values indicates that the proposed model is fit for a practical trial to detect appropriate face mask use in real time.

BACKGROUND AND OBJECTIVES: Photoplethysmography (PPG) is a device that measures the amount of light absorbed by the blood vessel, blood, and tissues, which can, in turn, translate into various measurements such as the variation in blood flow volume, heart rate variability, blood pressure, etc. Hence, PPG signals can produce a wide variety of biological information that can be useful for the detection and diagnosis of various health problems. In this review, we are interested in the possible health disorders that can be detected using PPG signals. METHODS: We applied PRISMA guidelines to systematically search various journal databases and identified 43 PPG studies that fit the criteria of this review. RESULTS: Twenty-five health issues were identified from these studies that were classified into six categories: cardiac, blood pressure, sleep health, mental health, diabetes, and miscellaneous. Various routes were employed in these PPG studies to perform the diagnosis: machine learning, deep learning, and statistical routes. The studies were reviewed and summarized. CONCLUSIONS: We identified limitations such as poor standardization of sampling frequencies and lack of publicly available PPG databases. We urge that future work should consider creating more publicly available databases so that a wide spectrum of health problems can be covered. We also want to promote the use of PPG signals as a potential precision medicine tool in both ambulatory and hospital settings.

Mental disorders (MDs) with onset in childhood or adolescence include neurodevelopmental disorders (NDDs) (intellectual disability and specific learning disabilities, such as dyslexia, attention deficit disorder (ADHD), and autism spectrum disorders (ASD)), as well as a broad range of mental health disorders (MHDs), including anxiety, depressive, stress-related and psychotic disorders. There is a high co-morbidity of NDDs and MHDs. Globally, there have been dramatic increases in the diagnosis of childhood-onset mental disorders, with a 2- to 3-fold rise in prevalence for several MHDs in the US over the past 20 years. Depending on the type of MD, children often grapple with social and communication deficits and difficulties adapting to changes in their environment, which can impact their ability to learn effectively. To improve outcomes for children, it is important to provide timely and effective interventions. This review summarises the range and effectiveness of AI-assisted tools, developed using machine learning models, which have been applied to address learning challenges in students with a range of NDDs. Our review summarises the evidence that AI tools can be successfully used to improve social interaction and supportive education. Based on the limitations of existing AI tools, we provide recommendations for the development of future AI tools with a focus on providing personalised learning for individuals with NDDs.

Automated sleep disorder detection is challenging because physiological symptoms can vary widely. These variations make it difficult to create effective sleep disorder detection models which support hu-man experts during diagnosis and treatment monitoring. From 2010 to 2021, authors of 95 scientific papers have taken up the challenge of automating sleep disorder detection. This paper provides an expert review of this work. We investigated whether digital technology and Artificial Intelligence (AI) can provide automated diagnosis support for sleep disorders. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines during the content discovery phase. We compared the performance of proposed sleep disorder detection methods, involving differ-ent datasets or signals. During the review, we found eight sleep disorders, of which sleep apnea and insomnia were the most studied. These disorders can be diagnosed using several kinds of biomedical signals, such as Electrocardiogram (ECG), Polysomnography (PSG), Electroencephalogram (EEG), Electromyogram (EMG), and snore sound. Subsequently, we established areas of commonality and distinctiveness. Common to all reviewed papers was that AI models were trained and tested with labelled physiological signals. Looking deeper, we discovered that 24 distinct algorithms were used for the detection task. The nature of these algorithms evolved, before 2017 only traditional Machine Learning (ML) was used. From 2018 onward, both ML and Deep Learning (DL) methods were used for sleep disorder detection. The strong emergence of DL algorithms has considerable implications for future detection systems because these algorithms demand significantly more data for training and testing when compared with ML. Based on our review results, we suggest that both type and amount of labelled data is crucial for the design of future sleep disorder detection systems because this will steer the choice of AI algorithm which establishes the desired decision support. As a guiding principle, more labelled data will help to represent the variations in symptoms. DL algorithms can extract information from these larger data quantities more effectively, therefore; we predict that the role of these algorithms will continue to expand.

Heart Rate Variability (HRV) is a good predictor of human health because the heart rhythm is modulated by a wide range of physiological processes. This statement embodies both challenges to and opportunities for HRV analysis. Opportunities arise from the wide-ranging applicability of HRV analysis for disease detection. The availability of modern high-quality sensors and the low data rate of heart rate signals make HRV easy to measure, communicate, store, and process. However, there are also significant obstacles that prevent a wider use of this technology. HRV signals are both nonstationary and nonlinear and, to the human eye, they appear noise-like. This makes them difficult to analyze and indeed the analysis findings are difficult to explain. Moreover, it is difficult to discriminate between the influences of different complex physiological processes on the HRV. These difficulties are compounded by the effects of aging and the presence of comorbidities. In this review, we have looked at scientific studies that have addressed these challenges with advanced signal processing and Artificial Intelligence (AI) methods.

Objectives. Fetal sex determination with ultrasound (US) examination is indicated in pregnancies at risk of X-linked genetic disorders or ambiguous genitalia. However, misdiagnoses often arise due to operator inexperience and technical difficulties while acquiring diagnostic images. We aimed to develop an efficient automated US-based fetal sex classification model that can facilitate efficient screening and reduce misclassification. Methods. We have developed a novel feature engineering model termed PFP-LHCINCA that employs pyramidal fixed-size patch generation with average pooling-based image decomposition, handcrafted feature extraction based on local phase quantization (LPQ), and histogram of oriented gradients (HOG) to extract directional and textural features and used Chi-square iterative neighborhood component analysis feature selection (CINCA), which iteratively selects the most informative feature vector for each image that minimizes calculated feature parameter-derived k-nearest neighbor-based misclassification rates. The model was trained and tested on a sizeable expert-labeled dataset comprising 339 males’ and 332 females’ fetal US images. One transverse fetal US image per subject zoomed to the genital area and standardized to 256 × 256 size was used for analysis. Fetal sex was annotated by experts on US images and confirmed postnatally. Results. Standard model performance metrics were compared using five shallow classifiers—k-nearest neighbor (kNN), decision tree, naïve Bayes, linear discriminant, and support vector machine (SVM)—with the hyperparameters tuned using a Bayesian optimizer. The PFP-LHCINCA model achieved a sex classification accuracy of ≥88% with all five classifiers and the best accuracy rates (>98%) with kNN and SVM classifiers. Conclusions. US-based fetal sex classification is feasible and accurate using the presented PFP-LHCINCA model. The salutary results support its clinical use for fetal US image screening for sex classification. The model architecture can be modified into deep learning models for training larger datasets.

Mental performance classification is a critical issue for brain-computer interfaces. Accurate and reliable classification of good or bad mental performance gives important clues for the preliminary diagnosis of some diseases and mental stress. In this work, we put forward an objective artificial intelligence model to quantify the clarity of thought during mental arithmetic tasks. The proposed model consists of: (i) multilevel feature extraction based on statistical and texture analysis methods, (ii) feature ranking and selection with a Chi2 method, (iii) classification, and (iv) weightless majority voting classifier. The novelty of the presented model comes from multilevel fused feature generation. The presented model was developed using 20 channel electroencephalography data from 36 subjects. The signals were captured while the subjects were performing mental arithmetic tasks. The individual datasets were labeled as either good or bad, based on the task results. We have obtained an accuracy of 96.77% using O2 channel with a k-nearest neighbor classifier and reached 100.0% accuracy with the majority voting classifier. Our results indicate that it is possible to determine mental performance with artificial intelligence. That might be a steppingstone to establish objective measures for the clarity of thought during a wide range of mental tasks.

Myocardial infarction (MI) is a life-threatening condition diagnosed acutely on the electrocardiogram (ECG). Several errors, such as noise, can impair the prediction of automated ECG diagnosis. Therefore, quantification and communication of model uncertainty are essential for reliable MI diagnosis.A Dirichlet DenseNet model that could analyze out-of-distribution data and detect misclassification of MI and normal ECG signals was developed. The DenseNet model was first trained with the pre-processed MI ECG signals (from the best lead V6) acquired from the Physikalisch-Technische Bundesanstalt (PTB) database, using the reverse Kullback-Leibler (KL) divergence loss. The model was then tested with newly synthesized ECG signals with added em and ma noise samples. Predictive entropy was used as an uncertainty measure to determine the misclassification of normal and MI signals. Model performance was evaluated using four uncertainty metrics: uncertainty sensitivity (UNSE), uncertainty specificity (UNSP), uncertainty accuracy (UNAC), and uncertainty precision (UNPR); the classification threshold was set at 0.3.The UNSE of the DenseNet model was low but increased over the studied decremental noise range (-6 to 24 dB), indicating that the model grew more confident in classifying the signals as they got less noisy. The model became more certain in its predictions from SNR values of 12 dB and 18 dB onwards, yielding UNAC values of 80% and 82.4% for em and ma noise signals, respectively. UNSP and UNPR values were close to 100% for em and ma noise signals, indicating that the model was self-aware of what it knew and didn't.Through this work, it has been established that the model is reliable as it was able to convey when it was not confident in the diagnostic information it was presenting. Thus, the model is trustworthy and can be used in healthcare applications, such as the emergency diagnosis of MI on ECGs.

Breast cancer is one of the largest single contributors to the burden of disease worldwide. Early detection of breast cancer has been shown to be associated with better overall clinical outcomes. Ultrasonography is a vital imaging modality in managing breast lesions. In addition, the development of computer-aided diagnosis (CAD) systems has further enhanced the importance of this imaging modality. Proper development of robust and reproducible CAD systems depends on the inclusion of different data from different populations and centers to considerate all variations in breast cancer pathology and minimize confounding factors. The current database contains ultrasound images and radiologist-defined masks of two sets of histologically proven benign and malignant lesions. Using this and similar pieces of data can aid in the development of robust CAD systems.

The gravity of ischemic stroke is the key factor in deciding upon the optimum therapeutic intervention. Ischemic stroke can be divided into three main groups: lacunar syndrome (LACS), partial anterior circulation syndrome (PACS), and total anterior circulation stroke (TACS), where the corresponding severity is mild, medium, and high, respectively. Herein, a unique method for the automatic detection of ischemic stroke severity is presented. The proposed system is based upon the extraction of higher order bispectrum entropy and its phase features from brain MRI (Magnetic Resonance Imaging) images. For classification, which is used to establish stroke severity, a support vector machine was incorporated into the design. The developed technique effectively detected the stroke lesion, and achieved a sensitivity, specificity, accuracy, and positive predictive value equal to 96.4%, 100%, 97.6% and 100%, respectively. The results were obtained without the need for manual intervention. This design is advantageous over state-of-the-art automated stroke severity detection systems, which require the reading neuroradiologist to manually determine the region of interest. Hence, the method is efficacious for delivering decision support in the diagnosis of ischemic stroke severity, thereby aiding the neuroradiologist in routine screening procedures.

Obstructive sleep apnea (OSA) is a common respiratory disorder marked by interruption of the respiratory tract and difficulty in breathing. The risk of serious health damage can be reduced if OSA is diagnosed and treated at an early stage. OSA is primarily diagnosed using polysomnography (PSG) monitoring performed for overnight sleep; furthermore, capturing PSG signals during the night is expensive, time-consuming, complex and highly inconvenient to patients. Hence, we are proposing to detect OSA automatically using respiratory and oximetry signals. The aim of this study is to develop a simple and computationally efficient wavelet-based automated system based on these signals to detect OSA in elderly subjects. In this study, we proposed an accurate, reliable, and less complex OSA automated detection system by using pulse oximetry (SpO

Early identification of coronary artery disease (CAD) can facilitate timely clinical intervention and save lives. This study aims to develop a machine learning framework that uses tensor analysis on heart rate (HR) signals to automate the CAD detection task. A third-order tensor representing a time-frequency relationship is constructed by fusing scalograms as vertical slices of the tensor. Each scalogram is computed from the considered time frame of a given HR signal. The derived scalogram represents the heterogeneity of data as a two-dimensional map. These two-dimensional maps are stacked one after the other horizontally along the z-axis to form a 3-way tensor for each HR signal. Each two-dimensional map is represented as a vertical slice in the xy - plane. Tensor factorization of such a fused tensor for every HR signal is performed using canonical polyadic (CP) decomposition. Only the core factor is retained later, excluding the three unitary matrices to provide the latent feature set for the detection task. The resultant latent features are then fed to machine learning classifiers for binary classification. Bayesian optimization is performed in a five-fold cross-validation strategy in search of the optimal machine learning classifier. The experimental results yielded the accuracy, sensitivity, and specificity of 96.62%, 96.53%, and 96.67%, respectively, with the bagged trees ensemble method. The proposed tensor decomposition deciphered higher-order interrelations among the considered time-frequency representations of HR signals.

BACKGROUND AND PURPOSE: Rotator cuff tear (RCT) and biceps tendinosis (BT) are the two most common shoulder disorders worldwide. These disorders can be diagnosed using magnetic resonance imaging (MRI), but the expert interpretation is manual, time-consuming, and subjected to human errors. Therefore, a fixed-size feature extraction model was created to objectively and accurately perform automated binary classification of RCT vs. normal and BT vs. normal on MRI images. MATERIALS AND METHODS: We have developed an exemplar deep feature extraction model to diagnose RCT and BT disorders. The model was tested on a new MR image dataset comprising transverse, sagittal, and coronal MRI images of the shoulder that had been organized into three cases. BT was studied on transverse MRI images (Case 1), while RCT was studied on sagittal (Case 2) and coronal MRI images (Case 3). Our model comprised deep feature generation using a pre-trained VGG19, feature selection using iterative neighborhood component analysis (INCA), and classification using shallow standard classifiers k-nearest neighbors (KNN), support vector machine (SVM), and artificial neural network (ANN). In the feature extraction phase, two fully connected layers were used to extract deep features from the original image, and sixteen fixed-size patches obtained by the division of the original image. This model was named Vision VGG19 (ViVGG), analogous to vision transformers (ViT). The feature vector is extracted from the raw image dataset, and 16 feature vectors are extracted from each fixed-size patch. Seventeen feature vectors obtained from each image are obtained from fc6 and fc7 layers of the pre-trained VGG19, are merged to obtain final feature vector. INCA was used to choose the top features from the created features, and the chosen features were classified using shallow classifiers. RESULTS: We defined three cases to evaluate the proposed ViVGG19 to diagnose RT and BCT disorders. Our proposed ViVGG19 model achieved more than 99% accuracy using the KNN classifier. CONCLUSIONS: ViVGG19 is a very effective model for detecting RCT and BT disorders on shoulder MRI images. The developed automated system is ready to be tested with a bigger diverse database obtained from different medical centers.

Recent advances in remote patient monitoring (RPM) systems can recognize various human activities to measure vital signs, including subtle motions from superficial vessels. There is a growing interest in applying artificial intelligence (AI) to this area of healthcare by addressing known limitations and challenges such as predicting and classifying vital signs and physical movements, which are considered crucial tasks. Federated learning is a relatively new AI technique designed to enhance data privacy by decentralizing traditional machine learning modeling. However, traditional federated learning requires identical architectural models to be trained across the local clients and global servers. This limits global model architecture due to the lack of local models heterogeneity. To overcome this, a novel federated learning architecture, FedStack, which supports ensembling heterogeneous architectural client models was proposed in this study. This work offers a protected privacy system for hospitalized in-patients in a decentralized approach and identifies optimum sensor placement. The proposed architecture was applied to a mobile health sensor benchmark dataset from 10 different subjects to classify 12 routine activities. Three AI models, ANN, CNN, and Bi-LSTM were trained on individual subject data. The federated learning architecture was applied to these models to build local and global models capable of state of the art performances. The local CNN model outperformed ANN and Bi-LSTM models on each subject data. Our proposed work has demonstrated better performance for heterogeneous stacking of the local models compared to homogeneous stacking. This work sets the stage to build an enhanced RPM system that incorporates client privacy to assist with clinical observations for patients in an acute mental health facility and ultimately help to prevent unexpected death., 47 Pages, Journal Article. Knowledge-Based Systems (2022)

Since the advent of COVID-19, the number of deaths has increased exponentially, boosting the requirement for various research studies that may correctly diagnose the illness at an early stage. Using chest X-rays, this study presents deep learning-based algorithms for classifying patients with COVID illness, healthy controls, and pneumonia classes. Data gathering, pre-processing, feature extraction, and classification are the four primary aspects of the approach. The pictures of chest X-rays utilized in this investigation came from various publicly available databases. The pictures were filtered to increase image quality in the pre-processing stage, and the chest X-ray images were de-noised using the empirical wavelet transform (EWT). Following that, four deep learning models were used to extract features. The first two models, Inception-V3 and Resnet-50, are based on transfer learning models. The Resnet-50 is combined with a temporal convolutional neural network (TCN) to create the third model. The fourth model is our suggested RESCOVIDTCNNet model, which integrates EWT, Resnet-50, and TCN. Finally, an artificial neural network (ANN) and a support vector machine were used to classify the data (SVM). Using five-fold cross-validation for 3-class classification, our suggested RESCOVIDTCNNet achieved a 99.5 percent accuracy. Our prototype can be utilized in developing nations where radiologists are in low supply to acquire a diagnosis quickly.

Coronary artery disease (CAD) is a prevalent disease with high morbidity and mortality rates. Invasive coronary angiography is the reference standard for diagnosing CAD but is costly and associated with risks. Noninvasive imaging like cardiac magnetic resonance (CMR) facilitates CAD assessment and can serve as a gatekeeper to downstream invasive testing. Machine learning methods are increasingly applied for automated interpretation of imaging and other clinical results for medical diagnosis. In this study, we proposed a novel CAD detection method based on CMR images by utilizing the feature extraction ability of deep neural networks and combining the features with the aid of a random forest for the very first time. It is necessary to convert image data to numeric features so that they can be used in the nodes of the decision trees. To this end, the predictions of multiple stand-alone convolutional neural networks (CNNs) were considered as input features for the decision trees. The capability of CNNs in representing image data renders our method a generic classification approach applicable to any image dataset. We named our method RF-CNN-F, which stands for Random Forest with CNN Features. We conducted experiments on a large CMR dataset that we have collected and made publicly accessible. Our method achieved excellent accuracy (99.18%) using Adam optimizer compared to a stand-alone CNN trained using fivefold cross validation (93.92%) tested on the same dataset.

Atrial fibrillation (AF) is a common arrhythmia that can lead to stroke, heart failure, and premature death. Manual screening of AF on electrocardiography (ECG) is time-consuming and prone to errors. To overcome these limitations, computer-aided diagnosis systems are developed using artificial intelligence techniques for automated detection of AF. Various machine learning and deep learning (DL) techniques have been developed for the automated detection of AF. In this review, we focused on the automated AF detection models developed using DL techniques. Twenty-four relevant articles published in international journals were reviewed. DL models based on deep neural network, convolutional neural network (CNN), recurrent neural network, long short-term memory, and hybrid structures were discussed. Our analysis showed that the majority of the studies used CNN models, which yielded the highest detection performance using ECG and heart rate variability signals. Details of the ECG databases used in the studies, performance metrics of the various models deployed, associated advantages and limitations, as well as proposed future work were summarized and discussed. This review paper serves as a useful resource for the researchers interested in developing innovative computer-assisted ECG-based DL approaches for AF detection.

Coronary artery disease is a major cause of morbidity and mortality worldwide. Its underlying histopathology is the atherosclerotic plaque, which comprises lipid, fibrous and—when chronic—calcium components. Intravascular ultrasound (IVUS) and intravascular optical coherence tomography (IVOCT) performed during invasive coronary angiography are reference standards for characterizing the atherosclerotic plaque. Fine image spatial resolution attainable with contemporary coronary computed tomographic angiography (CCTA) has enabled noninvasive plaque assessment, including identifying features associated with vulnerable plaques known to presage acute coronary events. Manual interpretation of IVUS, IVOCT and CCTA images demands scarce physician expertise and high time cost. This has motivated recent research into and development of artificial intelligence (AI)-assisted methods for image processing, feature extraction, plaque identification and characterization. We performed parallel searches of the medical and technical literature from 1995 to 2021 focusing respectively on human plaque characterization using various imaging modalities and the use of AI-assisted computer aided diagnosis (CAD) to detect and classify atherosclerotic plaques, including their composition and the presence of high-risk features denoting vulnerable plaques. A total of 122 publications were selected for evaluation and the analysis was summarized in terms of data sources, methods—machine versus deep learning—and performance metrics. Trends in AI-assisted plaque characterization are detailed and prospective research challenges discussed. Future directions for the development of accurate and efficient CAD systems to characterize plaque noninvasively using CCTA are proposed.

Cough-based disease detection is a hot research topic for machine learning, and much research has been published on the automatic detection of Covid-19. However, these studies are useful for the diagnosis of different diseases.In this work, we collected a new and large (n=642 subjects) cough sound dataset comprising four diagnostic categories: 'Covid-19', 'heart failure', 'acute asthma', and 'healthy', and used it to train, validate, and test a novel model designed for automatic detection.The model consists of four main components: novel feature generation based on a specifically directed knight pattern (DKP), signal decomposition using four pooling methods, feature selection using iterative neighborhood analysis (INCA), and classification using the k-nearest neighbor (kNN) classifier with ten-fold cross-validation. Multilevel multiple pooling decomposition combined with DKP yielded 41 feature vectors (40 extracted plus one original cough sound). From these, the ten best feature vectors were selected. Based on each vector's misclassification rate, redundant feature vectors were eliminated and then merged. The merged vector's most informative features automatically selected using INCA were input to a standard kNN classifier.The model, called DKPNet41, attained a high accuracy of 99.39% for cough sound-based multiclass classification of the four categories.The results obtained in the study showed that the DKPNet41 model automatically and efficiently classifies cough sounds for disease diagnosis.

Sleep apnea (SA) is characterized by intermittent episodes of apnea or hypopnea paused or reduced breathing, respectively each lasting at least ten seconds that occur during sleep. SA has an estimated global prevalence of 200 million and is associated with medical comorbidity, and sufferers are also more likely to sustain traffic- and work-related injury due to daytime somnolence. SA is amenable to treatment if detected early. Polysomnography (PSG) involving multi-channel signal acquisition is the reference standard for diagnosing SA but is onerous and costly. For home-based detection of SA, single-channel SpO 2 signal acquisition using portable pulse oximeters is feasible. Machine (ML) and deep learning (DL) models have been developed for automated classification of SA versus no SA using SpO 2 signals alone. In this work, we review studies published between 2012 and 2022 on the use of ML and DL for SpO 2 signal-based diagnosis of SA. A literature search based on PRISMA recommendations yielded 297 publications, of which 31 were selected after considering the inclusion and exclusion criteria. There were 20 ML and 11 DL models; their methods, differences, results, merits, and limitations were discussed. Many studies reported encouraging performance, which indicates the utility of SpO 2 signals in wearable devices for home-based SA detection.

Ultrasound (US) is an important imaging modality used to assess breast lesions for malignant features. In the past decade, many machine learning models have been developed for automated discrimination of breast cancer versus normal on US images, but few have classified the images based on the Breast Imaging Reporting and Data System (BI-RADS) classes. This work aimed to develop a model for classifying US breast lesions using a BI-RADS classification framework with a new multi-class US image dataset. We proposed a deep model that combined a novel pyramid triple deep feature generator (PTDFG) with transfer learning based on three pre-trained networks for creating deep features. Bilinear interpolation was applied to decompose the input image into four images of successively smaller dimensions, constituting a four-level pyramid for downstream feature generation with the pre-trained networks. Neighborhood component analysis was applied to the generated features to select each network's 1,000 most informative features, which were fed to support vector machine classifier for automated classification using a ten-fold cross-validation strategy. Our proposed model was validated using a new US image dataset containing 1,038 images divided into eight BI-RADS classes and histopathological results. We defined three classification schemes: Case 1 involved the classification of all images into eight categories; Case 2, classification of breast US images into five BI-RADS classes; and Case 3, classification of BI-RADS 4 lesions into benign versus malignant classes. Our PTDFG-based transfer learning model attained accuracy rates of 79.29%, 80.42%, and 88.67% for Case 1, Case 2, and Case 3, respectively.

Epilepsy is a disorder of the brain denoted by frequent seizures. The symptoms of seizure include confusion, abnormal staring, and rapid, sudden, and uncontrollable hand movements. Epileptic seizure detection methods involve neurological exams, blood tests, neuropsychological tests, and neuroimaging modalities. Among these, neuroimaging modalities have received considerable attention from specialist physicians. One method to facilitate the accurate and fast diagnosis of epileptic seizures is to employ computer-aided diagnosis systems (CADS) based on deep learning (DL) and neuroimaging modalities. This paper has studied a comprehensive overview of DL methods employed for epileptic seizures detection and prediction using neuroimaging modalities. First, DL-based CADS for epileptic seizures detection and prediction using neuroimaging modalities are discussed. Also, descriptions of various datasets, preprocessing algorithms, and DL models which have been used for epileptic seizures detection and prediction have been included. Then, research on rehabilitation tools has been presented, which contains brain-computer interface (BCI), cloud computing, internet of things (IoT), hardware implementation of DL techniques on field-programmable gate array (FPGA), etc. In the discussion section, a comparison has been carried out between research on epileptic seizure detection and prediction. The challenges in epileptic seizures detection and prediction using neuroimaging modalities and DL models have been described. In addition, possible directions for future works in this field, specifically for solving challenges in datasets, DL, rehabilitation, and hardware models, have been proposed. The final section is dedicated to the conclusion which summarizes the significant findings of the paper.

OBJECTIVE: Parkinson's disease (PD) is a common neurological disorder with variable clinical manifestations and magnetic resonance imaging (MRI) findings. We propose a handcrafted image classification model that can accurately (i) classify different PD stages, (ii) detect comorbid dementia, and (iii) discriminate PD-related motor symptoms. METHODS: Selected image datasets from three PD studies were used to develop the classification model. Our proposed novel automated system was developed in four phases: (i) texture features are extracted from the non-fixed size patches. In the feature extraction phase, a pyramid histogram-oriented gradient (PHOG) image descriptor is used. (ii) In the feature selection phase, four feature selectors: neighborhood component analysis (NCA), Chi2, minimum redundancy maximum relevancy (mRMR), and ReliefF are used to generate four feature vectors. (iii) Two classifiers: k-nearest neighbor (kNN) and support vector machine (SVM) are used in the classification step. A ten-fold cross-validation technique is used to validate the results. (iv) Eight predicted vectors are generated using four selected feature vectors and two classifiers. Finally, iterative majority voting (IMV) is used to attain general classification results. Therefore, this model is named nested patch-PHOG-multiple feature selectors and multiple classifiers-IMV (NP-PHOG-MFSMCIMV). RESULTS: Our presented NP-PHOG-MFSMCIMV model achieved 99.22, 98.70, and 99.53% accuracies for the collected PD stages, PD dementia, and PD symptoms classification datasets, respectively. SIGNIFICANCE: The obtained accuracies (over 98% for all states) demonstrated the performance of developed NP-PHOG-MFSMCIMV model in automated PD state classification.

Breast histology image classification is a crucial step in the early diagnosis of breast cancer. In breast pathological diagnosis, Convolutional Neural Networks (CNNs) have demonstrated great success using digitized histology slides. However, tissue classification is still challenging due to the high visual variability of the large-sized digitized samples and the lack of contextual information. In this paper, we propose a novel CNN, called Multi-level Context and Uncertainty aware (MCUa) dynamic deep learning ensemble model.MCUamodel consists of several multi-level context-aware models to learn the spatial dependency between image patches in a layer-wise fashion. It exploits the high sensitivity to the multi-level contextual information using an uncertainty quantification component to accomplish a novel dynamic ensemble model.MCUamodelhas achieved a high accuracy of 98.11% on a breast cancer histology image dataset. Experimental results show the superior effectiveness of the proposed solution compared to the state-of-the-art histology classification models., accepted by IEEE Transactions on Biomedical Engineering

BACKGROUND: Human emotions greatly affect the actions of a person. The automated emotion recognition has applications in multiple domains such as health care, e-learning, surveillance, etc. The development of computer-aided diagnosis (CAD) tools has led to the automated recognition of human emotions.OBJECTIVE: This review paper provides an insight into various methods employed using electroencephalogram (EEG), facial, and speech signals coupled with multi-modal emotion recognition techniques. In this work, we have reviewed most of the state-of-the-art papers published on this topic.METHOD: This study was carried out by considering the various emotion recognition (ER) models proposed between 2016 and 2021. The papers were analysed based on methods employed, classifier used and performance obtained.RESULTS: There is a significant rise in the application of deep learning techniques for ER. They have been widely applied for EEG, speech, facial expression, and multimodal features to develop an accurate ER model.CONCLUSION: Our study reveals that most of the proposed machine and deep learning-based systems have yielded good performances for automated ER in a controlled environment. However, there is a need to obtain high performance for ER even in an uncontrolled environment.

This paper presents a scientific foundation for automated stroke severity classification. We have constructed and assessed a system which extracts diagnostically relevant information from Magnetic Resonance Imaging (MRI) images. The design was based on 267 images that show the brain from individual subjects after stroke. They were labeled as either Lacunar Syndrome (LACS), Partial Anterior Circulation Syndrome (PACS), or Total Anterior Circulation Stroke (TACS). The labels indicate different physiological processes which manifest themselves in distinct image texture. The processing system was tasked with extracting texture information that could be used to classify a brain MRI image from a stroke survivor into either LACS, PACS, or TACS. We analyzed 6475 features that were obtained with Gray-Level Run Length Matrix (GLRLM), Higher Order Spectra (HOS), as well as a combination of Discrete Wavelet Transform (DWT) and Gray-Level Co-occurrence Matrix (GLCM) methods. The resulting features were ranked based on the p-value extracted with the Analysis Of Variance (ANOVA) algorithm. The ranked features were used to train and test four types of Support Vector Machine (SVM) classification algorithms according to the rules of 10-fold cross-validation. We found that SVM with Radial Basis Function (RBF) kernel achieves: Accuracy (ACC) = 93.62%, Specificity (SPE) = 95.91%, Sensitivity (SEN) = 92.44%, and Dice-score = 0.95. These results indicate that computer aided stroke severity diagnosis support is possible. Such systems might lead to progress in stroke diagnosis by enabling healthcare professionals to improve diagnosis and management of stroke patients with the same resources.

Sleep is highly essential for maintaining metabolism of the body and mental balance for increased productivity and concentration. Often, sleep is analyzed using macrostructure sleep stages which alone cannot provide information about the functional structure and stability of sleep. The cyclic alternating pattern (CAP) is a physiological recurring electroencephalogram (EEG) activity occurring in the brain during sleep and captures microstructure of the sleep and can be used to identify sleep instability. The CAP can also be associated with various sleep-related pathologies, and can be useful in identifying various sleep disorders. Conventionally, sleep is analyzed using polysomnogram (PSG) in various sleep laboratories by trained physicians and medical practitioners. However, PSG-based manual sleep analysis by trained medical practitioners is onerous, tedious and unfavourable for patients. Hence, a computerized, simple and patient convenient system is highly desirable for monitoring and analysis of sleep. In this study, we have proposed a system for automated identification of CAP phase-A and phase-B. To accomplish the task, we have utilized the openly accessible CAP sleep database. The study is performed using two single-channel EEG modalities and their combination. The model is developed using EEG signals of healthy subjects as well as patients suffering from six different sleep disorders namely nocturnal frontal lobe epilepsy (NFLE), sleep-disordered breathing (SDB), narcolepsy, periodic leg movement disorder (PLM), insomnia and rapid eye movement behavior disorder (RBD) subjects. An optimal orthogonal wavelet filter bank is used to perform the wavelet decomposition and subsequently, entropy and Hjorth parameters are extracted from the decomposed coefficients. The extracted features have been applied to different machine learning algorithms. The best performance is obtained using ensemble of bagged tress (EBagT) classifier. The proposed method has obtained the average classification accuracy of 84%, 83%, 81%, 78%, 77%, 76% and 72% for NFLE, healthy, SDB, narcolepsy, PLM, insomnia and RBD subjects, respectively in discriminating phases A and B using a balanced database. Our developed model yielded an average accuracy of 78% when all 77 subjects including healthy and sleep disordered patients are considered. Our proposed system can assist the sleep specialists in an automated and efficient analysis of sleep using sleep microstructure.

Kidney stone is a commonly seen ailment and is usually detected by urologists using computed tomography (CT) images. It is difficult and time-consuming to detect small stones in CT images. Hence, an automated system can help clinicians to detect kidney stones accurately. In this work, a novel transfer learning-based image classification method (ExDark19) has been proposed to detect kidney stones using CT images. The iterative neighborhood component analysis (INCA) is employed to select the most informative feature vectors and these selected features vectors are fed to the k nearest neighbor (kNN) classifier to detect kidney stones with a ten-fold cross-validation (CV) strategy. The proposed ExDark19 model yielded an accuracy of 99.22% with 10-fold CV and 99.71% using the hold-out validation method. Our results demonstrate that the proposed ExDark19 detect kidney stones over 99% accuracies for two validation techniques. This developed automated system can assist the urologists to validate their manual screening of kidney stones and hence reduce the possible human error.

Background and objective Parkinson's disease (PD) is one of the most common diseases worldwide which reduces quality of life of patients and their family members. The electroencephalogram (EEG) signals coupled with various advanced machine-learning algorithms have been widely used to detect PD automatically. In this paper, we propose a novel aspirin pattern to detect PD accurately using EEG signals. Method In this research, the feature generation ability of a chemical graph is investigated. Therefore, this work presents a new graph-based aspirin model for automated PD detection using EEG signals. The proposed method consists of (i) multilevel feature generation phase involving new aspirin pattern, statistical moments, and maximum absolute pooling (MAP), (ii) selection of most discriminative features using neighborhood component analysis (NCA), and (iii) classification using k nearest neighbor (kNN) for automated detection of PD and (iv) iterative majority voting. Results A public dataset has been used to develop the proposed model. Two cases are created, and these cases consisted of two classes. Leave one subject out (LOSO) validation have been used to calculate robust results. Our proposal achieved 93.57% and 95.48% classification accuracies for Case 1 and Case 2 respectively. Conclusion Our developed automated PD model is accurate and equipped to be tested with more diverse EEG datasets.

A sleep disorder is a medical condition that affects an individual’s regular sleeping pattern and routine, hence negatively affecting the individual’s health. The traditional procedures of identifying sleep disorders by clinicians involve questionnaires and polysomnography (PSG), which are subjective, time-consuming, and inconvenient. Hence, an automated sleep disorder identification is required to overcome these limitations. In the proposed study, we have proposed a method using electroencephalogram (EEG) signals for the automated identification of six sleep disorders, namely insomnia, nocturnal frontal lobe epilepsy (NFLE), narcolepsy, rapid eye movement behavior disorder (RBD), periodic leg movement disorder (PLM), and sleep-disordered breathing (SDB). To the best of our belief, this is one of the first studies ever undertaken to identify sleep disorders using EEG signals employing cyclic alternating pattern (CAP) sleep database. After sleep-scoring EEG epochs, we have created eight different data subsets of EEG epochs to develop the proposed model. A novel optimal triplet half-band filter bank (THFB) is used to obtain the subbands of EEG signals. We have extracted Hjorth parameters from subbands of EEG epochs. The selected features are fed to various supervised machine learning algorithms for the automated classification of sleep disorders. Our proposed system has obtained the highest accuracy of 99.2%, 98.2%, 96.2%, 98.3%, 98.8%, and 98.8% for insomnia, narcolepsy, NFLE, PLM, RBD, and SDB classes against normal healthy subjects, respectively, applying ensemble boosted trees classifier. As a result, we have attained the highest accuracy of 91.3% to identify the type of sleep disorder. The proposed method is simple, fast, efficient, and may reduce the challenges faced by medical practitioners during the diagnosis of various sleep disorders accurately in less time at sleep clinics and homes.

Radiomics is a newcomer field that has opened new windows for precision medicine. It is related to extraction of a large number of quantitative features from medical images, which may be difficult to detect visually. Underlying tumor biology can change physical properties of tissues, which affect patterns of image pixels and radiomics features. The main advantage of radiomics is that it can characterize the whole tumor non-invasively, even after a single sampling from an image. Therefore, it can be linked to a "digital biopsy". Physicians need to know about radiomics features to determine how their values correlate with the appearance of lesions and diseases. Indeed, physicians need practical references to conceive of basics and concepts of each radiomics feature without knowing their sophisticated mathematical formulas. In this review, commonly used radiomics features are illustrated with practical examples to help physicians in their routine diagnostic procedures.

Artificial Intelligence in healthcare employs machine learning algorithms to emulate human cognition in the analysis of complicated or large sets of data. Specifically, artificial intelligence taps on the ability of computer algorithms and software with allowable thresholds to make deterministic approximate conclusions. In comparison to traditional technologies in healthcare, artificial intelligence enhances the process of data analysis without the need for human input, producing nearly equally reliable, well defined output. Schizophrenia is a chronic mental health condition that affects millions worldwide, with impairment in thinking and behaviour that may be significantly disabling to daily living. Multiple artificial intelligence and machine learning algorithms have been utilized to analyze the different components of schizophrenia, such as in prediction of disease, and assessment of current prevention methods. These are carried out in hope of assisting with diagnosis and provision of viable options for individuals affected. In this paper, we review the progress of the use of artificial intelligence in schizophrenia.