Optimised stacked machine learning algorithms for genomics and genetics disorder detection in the healthcare industry.

With recent advances in precision medicine and healthcare computing, there is an enormous demand for developing machine learning algorithms in genomics to enhance the rapid analysis of disease disorders. Technological advancement in genomics and imaging provides clinicians with enormous amounts of data, but prediction is still mostly subjective, resulting in problematic medical treatment. Machine learning is being employed in several domains of the healthcare sector, encompassing clinical research, early disease identification, and medicinal innovation with a historical perspective. The main objective of this study is to detect patients who, based on several medical standards, are more susceptible to having a genetic disorder. A genetic disease prediction algorithm was employed, leveraging the patient's health history to evaluate the probability of diagnosing a genetic disorder. We developed a computationally efficient machine learning approach to predict the overall lifespan of patients with a genomics disorder and to classify and predict patients with a genetic disease. The SVM, RF, and ETC are stacked using two-layer meta-estimators to develop the proposed model. The first layer comprises all the baseline models employed to predict the outcomes based on the dataset. The second layer comprises a component known as a meta-classifier. Results from the experiment indicate that the model achieved an accuracy of 90.45% and a recall score of 90.19%. The area under the curve (AUC) for mitochondrial diseases is 98.1%; for multifactorial diseases, it is 97.5%; and for single-gene inheritance, it is 98.8%. The proposed approach presents a novel method for predicting patient prognosis in a manner that is unbiased, accurate, and comprehensive. The proposed approach outperforms human professionals using the current clinical standard for genetic disease classification in terms of identification accuracy. The implementation of stacked will significantly improve the field of biomedical research by improving the anticipation of genetic diseases.

Attention-based deep learning framework to recognize diabetes disease from cellular retinal images.

A medical disorder known as diabetic retinopathy (DR) affects people who suffer from diabetes. Many people are visually impaired due to DR. Primary cause of DR in patients is high blood sugar, and it affects blood vessels available in the retinal cell. The recent advancement in deep learning and computer vision methods, and their automation applications can recognize the presence of DR in retinal cells and vessel images. Authors have proposed an attention-based hybrid model to recognize diabetes in early stage to prevent harmful clauses. Proposed methodology uses DenseNet121 architecture for convolution learning and then, the feature vector will be enhanced with channel and spatial attention model. The proposed architecture also simulates binary and multiclass classification to recognize the infection and the spreading of disease. Binary classification recognizes DR images either positive or negative, while multiclass classification represents an infection on a scale of 0-4. Simulation of the proposed methodology has achieved 98.57% and 99.01% accuracy for multiclass and binary classification, respectively. Simulation of the study also explored the impact of data augmentation to make the proposed model robust and generalized. Attention-based deep learning model has achieved remarkable accuracy to detect diabetic infection from retinal cellular images.

Cloud-edge load balancing distributed protocol for IoE services using swarm intelligence.

Rapid development of the Internet of Everything (IoE) and cloud services offer a vital role in the growth of smart applications. It provides scalability with the collaboration of cloud servers and copes with a big amount of collected data for network systems. Although, edge computing supports efficient utilization of communication bandwidth, and latency requirements to facilitate smart embedded systems. However, it faces significant research issues regarding data aggregation among heterogeneous network services and objects. Moreover, distributed systems are more precise for data access and storage, thus machine-to-machine is needed to be secured from unpredictable events. As a result, this research proposed secured data management with distributed load balancing protocol using particle swarm optimization, which aims to decrease the response time for cloud users and effectively maintain the integrity of network communication. It combines distributed computing and shift high cost computations closer to the requesting node to reduce latency and transmission overhead. Moreover, the proposed work also protects the communicating machines from malicious devices by evaluating the trust in a controlled manner. Simulation results revealed a significant performance of the proposed protocol in comparison to other solutions in terms of energy consumption by 20%, success rate by 17%, end-to-end delay by 14%, and network cost by 19% as average in the light of various performance metrics.

Device-to-Device (D2D) Multi-Criteria Learning Algorithm Using Secured Sensors.

Wireless networks and the Internet of things (IoT) have proven rapid growth in the development and management of smart environments. These technologies are applied in numerous research fields, such as security surveillance, Internet of vehicles, medical systems, etc. The sensor technologies and IoT devices are cooperative and allow the collection of unpredictable factors from the observing field. However, the constraint resources of distributed battery-powered sensors decrease the energy efficiency of the IoT network and increase the delay in receiving the network data on users' devices. It is observed that many solutions are proposed to overcome the energy deficiency in smart applications; though, due to the mobility of the nodes, lots of communication incurs frequent data discontinuity, compromising the data trust. Therefore, this work introduces a D2D multi-criteria learning algorithm for IoT networks using secured sensors, which aims to improve the data exchange without imposing additional costs and data diverting for mobile sensors. Moreover, it reduces the compromising threats in the presence of anonymous devices and increases the trustworthiness of the IoT-enabled communication system with the support of machine learning. The proposed work was tested and analyzed using broad simulation-based experiments and demonstrated the significantly improved performance of the packet delivery ratio by 17%, packet disturbances by 31%, data delay by 22%, energy consumption by 24%, and computational complexity by 37% for realistic network configurations.

Sustainable Data-Driven Secured Optimization Using Dynamic Programming for Green Internet of Things.

The development of smart applications has benefited greatly from the expansion of wireless technologies. A range of tasks are performed, and end devices are made capable of communicating with one another with the support of artificial intelligence technology. The Internet of Things (IoT) increases the efficiency of communication networks due to its low costs and simple management. However, it has been demonstrated that many systems still need an intelligent strategy for green computing. Establishing reliable connectivity in Green-IoT (G-IoT) networks is another key research challenge. With the integration of edge computing, this study provides a Sustainable Data-driven Secured optimization model (SDS-GIoT) that uses dynamic programming to provide enhanced learning capabilities. First, the proposed approach examines multi-variable functions and delivers graph-based link predictions to locate the optimal nodes for edge networks. Moreover, it identifies a sub-path in multistage to continue data transfer if a route is unavailable due to certain communication circumstances. Second, while applying security, edge computing provides offloading services that lower the amount of processing power needed for low-constraint nodes. Finally, the SDS-GIoT model is verified with various experiments, and the performance results demonstrate its significance for a sustainable environment against existing solutions.

An Optimization Model with Network Edges for Multimedia Sensors Using Artificial Intelligence of Things.

In modern years, network edges have been explored by many applications to lower communication and management costs. They are also integrated with the internet of things (IoT) to achieve network design, in terms of scalability and heterogeneous services for multimedia applications. Many proposed solutions are performing a vital role in the development of robust protocols and reducing the response time for critical networks. However, most of them are not able to support the forwarding processes of high multimedia traffic under dynamic characteristics with constraint bandwidth. Moreover, they increase the rate of data loss in an uncertain environment and compromise network performance by increasing delivery delay. Therefore, this paper presents an optimization model with mobile edges for multimedia sensors using artificial intelligence of things, which aims to maintain the process of real-time data collection with low consumption of resources. Moreover, it improves the unpredictability of network communication with the integration of software-defined networks (SDN) and mobile edges. Firstly, it utilizes the artificial intelligence of things (AIoT), forming the multi-hop network and guaranteed the primary services for constraints network with stable resources management. Secondly, the SDN performs direct association with mobile edges to support the load balancing for multimedia sensors and centralized the management. Finally, multimedia traffic is heading towards applications in an unchanged form and without negotiating using the sharing of subkeys. The experimental results demonstrated its effectiveness for delivery rate by an average of 35%, processing delay by an average of 29%, network overheads by an average of 41%, packet drop ratio by an average of 39%, and packet retransmission by an average of 34% against existing solutions.

Prediction of COVID-19 - Pneumonia based on Selected Deep Features and One Class Kernel Extreme Learning Machine.

In this work, we propose a deep learning framework for the classification of COVID-19 pneumonia infection from normal chest CT scans. In this regard, a 15-layered convolutional neural network architecture is developed which extracts deep features from the selected image samples - collected from the Radiopeadia. Deep features are collected from two different layers, global average pool and fully connected layers, which are later combined using the max-layer detail (MLD) approach. Subsequently, a Correntropy technique is embedded in the main design to select the most discriminant features from the pool of features. One-class kernel extreme learning machine classifier is utilized for the final classification to achieving an average accuracy of 95.1%, and the sensitivity, specificity & precision rate of 95.1%, 95%, & 94% respectively. To further verify our claims, detailed statistical analyses based on standard error mean (SEM) is also provided, which proves the effectiveness of our proposed prediction design.

Deep Learning-Based COVID-19 Detection Using CT and X-Ray Images: Current Analytics and Comparisons.

Currently, the world faces a novel coronavirus disease 2019 (COVID-19) challenge and infected cases are increasing exponentially. COVID-19 is a disease that has been reported by the WHO in March 2020, caused by a virus called the SARS-CoV-2. As of 10 March 2021, more than 150 million people were infected and 3v million died. Researchers strive to find out about the virus and recommend effective actions. An unprecedented increase in pathogens is happening and a major attempt is being made to tackle the epidemic. This article presents deep learning-based COVID-19 detection using CT and X-ray images and data analytics on its spread worldwide. This article's research structure builds on a recent analysis of the COVID-19 data and prospective research to systematize current resources, help the researchers, practitioners by using in-depth learning methodologies to build solutions for the COVID-19 pandemic.

Real-Time Diagnosis System of COVID-19 Using X-Ray Images and Deep Learning.

The novel coronavirus named COVID-19 has quickly spread among humans worldwide, and the situation remains hazardous to the health system. The existence of this virus in the human body is identified through sputum or blood samples. Furthermore, computed tomography (CT) or X-ray has become a significant tool for quick diagnoses. Thus, it is essential to develop an online and real-time computer-aided diagnosis (CAD) approach to support physicians and avoid further spreading of the disease. In this research, a convolutional neural network (CNN) -based Residual neural network (ResNet50) has been employed to detect COVID-19 through chest X-ray images and achieved 98% accuracy. The proposed CAD system will receive the X-ray images from the remote hospitals/healthcare centers and perform diagnostic processes. Furthermore, the proposed CAD system uses advanced load balancer and resilience features to achieve fault tolerance with zero delays and perceives more infected cases during this pandemic.

A comprehensive study of mobile-health based assistive technology for the healthcare of dementia and Alzheimer's disease (AD).

Assistive technology (AT) involvement in therapeutic treatment has provided simple and efficient healthcare solutions to people. Within a short span of time, mobile health (mHealth) has grown rapidly for assisting people living with a chronic disorder. This research paper presents the comprehensive study to identify and review existing mHealth dementia applications (apps), and also synthesize the evidence of using these applications in assisting people with dementia including Alzheimer's disease (AD) and their caregivers. Six electronic databases searched with the purpose of finding literature-based evidence. The search yielded 2818 research articles, with 29 meeting quantified inclusion and exclusion criteria. Six groups and their associated sub-groups emerged from the literature. The main groups are (1) activities of daily living (ADL) based cognitive training, (2) monitoring, (3) dementia screening, (4) reminiscence and socialization, (5) tracking, and (6) caregiver support. Moreover, two commercial mobile application stores i.e., Apple App Store (iOS) and Google Play Store (Android) explored with the intention of identifying the advantages and disadvantages of existing commercially available dementia and AD healthcare apps. From 678 apps, a total of 38 mobile apps qualified as per defined exclusion and inclusion criteria. The shortlisted commercial apps generally targeted different aspects of dementia as identified in research articles. This comprehensive study determined the feasibility of using mobile Health based applications for dementia including AD individuals and their caregivers regardless of limited available research, and these apps have capability to incorporate a variety of strategies and resources to dementia community care.

Detecting Pneumonia using Convolutions and Dynamic Capsule Routing for Chest X-ray Images.

An entity's existence in an image can be depicted by the activity instantiation vector from a group of neurons (called capsule). Recently, multi-layered capsules, called CapsNet, have proven to be state-of-the-art for image classification tasks. This research utilizes the prowess of this algorithm to detect pneumonia from chest X-ray (CXR) images. Here, an entity in the CXR image can help determine if the patient (whose CXR is used) is suffering from pneumonia or not. A simple model of capsules (also known as Simple CapsNet) has provided results comparable to best Deep Learning models that had been used earlier. Subsequently, a combination of convolutions and capsules is used to obtain two models that outperform all models previously proposed. These models-Integration of convolutions with capsules (ICC) and Ensemble of convolutions with capsules (ECC)-detect pneumonia with a test accuracy of 95.33% and 95.90%, respectively. The latter model is studied in detail to obtain a variant called EnCC, where n = 3, 4, 8, 16. Here, the E4CC model works optimally and gives test accuracy of 96.36%. All these models had been trained, validated, and tested on 5857 images from Mendeley.

Machine Learning Algorithms and Fault Detection for Improved Belief Function Based Decision Fusion in Wireless Sensor Networks.

Decision fusion is used to fuse classification results and improve the classification accuracy in order to reduce the consumption of energy and bandwidth demand for data transmission. The decentralized classification fusion problem was the reason to use the belief function-based decision fusion approach in Wireless Sensor Networks (WSNs). With the consideration of improving the belief function fusion approach, we have proposed four classification techniques, namely Enhanced K-Nearest Neighbor (EKNN), Enhanced Extreme Learning Machine (EELM), Enhanced Support Vector Machine (ESVM), and Enhanced Recurrent Extreme Learning Machine (ERELM). In addition, WSNs are prone to errors and faults because of their different software, hardware failures, and their deployment in diverse fields. Because of these challenges, efficient fault detection methods must be used to detect faults in a WSN in a timely manner. We have induced four types of faults: offset fault, gain fault, stuck-at fault, and out of bounds fault, and used enhanced classification methods to solve the sensor failure issues. Experimental results show that ERELM gave the first best result for the improvement of the belief function fusion approach. The other three proposed techniques ESVM, EELM, and EKNN provided the second, third, and fourth best results, respectively. The proposed enhanced classifiers are used for fault detection and are evaluated using three performance metrics, i.e., Detection Accuracy (DA), True Positive Rate (TPR), and Error Rate (ER). Simulations show that the proposed methods outperform the existing techniques and give better results for the belief function and fault detection in WSNs.

Fault Detection in Wireless Sensor Networks through the Random Forest Classifier.

Wireless Sensor Networks (WSNs) are vulnerable to faults because of their deployment in unpredictable and hazardous environments. This makes WSN prone to failures such as software, hardware, and communication failures. Due to the sensor's limited resources and diverse deployment fields, fault detection in WSNs has become a daunting task. To solve this problem, Support Vector Machine (SVM), Convolutional Neural Network (CNN), Stochastic Gradient Descent (SGD), Multilayer Perceptron (MLP), Random Forest (RF), and Probabilistic Neural Network (PNN) classifiers are used for classification of gain, offset, spike, data loss, out of bounds, and stuck-at faults at the sensor level. Out of six faults, two of them are induced in the datasets, i.e., spike and data loss faults. The results are compared on the basis of their Detection Accuracy (DA), True Positive Rate (TPR), Matthews Correlation Coefficients (MCC), and F1-score. In this paper, a comparative analysis is performed among the classifiers mentioned previously on real-world datasets. Simulations show that the RF algorithm secures a better fault detection rate than the rest of the classifiers.

Lung Nodule Detection based on Ensemble of Hand Crafted and Deep Features.

Lung cancer is considered as a deadliest disease worldwide due to which 1.76 million deaths occurred in the year 2018. Keeping in view its dreadful effect on humans, cancer detection at a premature stage is a more significant requirement to reduce the probability of mortality rate. This manuscript depicts an approach of finding lung nodule at an initial stage that comprises of three major phases: (1) lung nodule segmentation using Otsu threshold followed by morphological operation; (2) extraction of geometrical, texture and deep learning features for selecting optimal features; (3) The optimal features are fused serially for classification of lung nodule into two categories that is malignant and benign. The lung image database consortium image database resource initiative (LIDC-IDRI) is used for experimentation. The experimental outcomes show better performance of presented approach as compared with the existing methods.

Region Extraction and Classification of Skin Cancer: A Heterogeneous framework of Deep CNN Features Fusion and Reduction.

Cancer is one of the leading causes of deaths in the last two decades. It is either diagnosed malignant or benign - depending upon the severity of the infection and the current stage. The conventional methods require a detailed physical inspection by an expert dermatologist, which is time-consuming and imprecise. Therefore, several computer vision methods are introduced lately, which are cost-effective and somewhat accurate. In this work, we propose a new automated approach for skin lesion detection and recognition using a deep convolutional neural network (DCNN). The proposed cascaded design incorporates three fundamental steps including; a) contrast enhancement through fast local Laplacian filtering (FlLpF) along HSV color transformation; b) lesion boundary extraction using color CNN approach by following XOR operation; c) in-depth features extraction by applying transfer learning using Inception V3 model prior to feature fusion using hamming distance (HD) approach. An entropy controlled feature selection method is also introduced for the selection of the most discriminant features. The proposed method is tested on PH2 and ISIC 2017 datasets, whereas the recognition phase is validated on PH2, ISBI 2016, and ISBI 2017 datasets. From the results, it is concluded that the proposed method outperforms several existing methods and attained accuracy 98.4% on PH2 dataset, 95.1% on ISBI dataset and 94.8% on ISBI 2017 dataset.

A New Approach for Brain Tumor Segmentation and Classification Based on Score Level Fusion Using Transfer Learning.

Brain tumor is one of the most death defying diseases nowadays. The tumor contains a cluster of abnormal cells grouped around the inner portion of human brain. It affects the brain by squeezing/ damaging healthy tissues. It also amplifies intra cranial pressure and as a result tumor cells growth increases rapidly which may lead to death. It is, therefore desirable to diagnose/ detect brain tumor at an early stage that may increase the patient survival rate. The major objective of this research work is to present a new technique for the detection of tumor. The proposed architecture accurately segments and classifies the benign and malignant tumor cases. Different spatial domain methods are applied to enhance and accurately segment the input images. Moreover Alex and Google networks are utilized for classification in which two score vectors are obtained after the softmax layer. Further, both score vectors are fused and supplied to multiple classifiers along with softmax layer. Evaluation of proposed model is done on top medical image computing and computer-assisted intervention (MICCAI) challenge datasets i.e., multimodal brain tumor segmentation (BRATS) 2013, 2014, 2015, 2016 and ischemic stroke lesion segmentation (ISLES) 2018 respectively.

A Deep Learning Approach for Automated Diagnosis and Multi-Class Classification of Alzheimer's Disease Stages Using Resting-State fMRI and Residual Neural Networks.

Alzheimer's disease (AD) is an incurable neurodegenerative disorder accounting for 70%-80% dementia cases worldwide. Although, research on AD has increased in recent years, however, the complexity associated with brain structure and functions makes the early diagnosis of this disease a challenging task. Resting-state functional magnetic resonance imaging (rs-fMRI) is a neuroimaging technology that has been widely used to study the pathogenesis of neurodegenerative diseases. In literature, the computer-aided diagnosis of AD is limited to binary classification or diagnosis of AD and MCI stages. However, its applicability to diagnose multiple progressive stages of AD is relatively under-studied. This study explores the effectiveness of rs-fMRI for multi-class classification of AD and its associated stages including CN, SMC, EMCI, MCI, LMCI, and AD. A longitudinal cohort of resting-state fMRI of 138 subjects (25 CN, 25 SMC, 25 EMCI, 25 LMCI, 13 MCI, and 25 AD) from Alzheimer's Disease Neuroimaging Initiative (ADNI) is studied. To provide a better insight into deep learning approaches and their applications to AD classification, we investigate ResNet-18 architecture in detail. We consider the training of the network from scratch by using single-channel input as well as performed transfer learning with and without fine-tuning using an extended network architecture. We experimented with residual neural networks to perform AD classification task and compared it with former research in this domain. The performance of the models is evaluated using precision, recall, f1-measure, AUC and ROC curves. We found that our networks were able to significantly classify the subjects. We achieved improved results with our fine-tuned model for all the AD stages with an accuracy of 100%, 96.85%, 97.38%, 97.43%, 97.40% and 98.01% for CN, SMC, EMCI, LMCI, MCI, and AD respectively. However, in terms of overall performance, we achieved state-of-the-art results with an average accuracy of 97.92% and 97.88% for off-the-shelf and fine-tuned models respectively. The Analysis of results indicate that classification and prediction of neurodegenerative brain disorders such as AD using functional magnetic resonance imaging and advanced deep learning methods is promising for clinical decision making and have the potential to assist in early diagnosis of AD and its associated stages.

An automated nuclei segmentation of leukocytes from microscopic digital images.

Leukemia is a life-threatening disease. So far diagnosing of leukemia is manually carried out by the Hematologists that is time-consuming and error-prone. The crucial problem is leukocytes' nuclei segmentation precisely. This paper presents a novel technique to solve the problem by applying statistical methods of Gaussian mixture model through expectation maximization for the basic and challenging step of leukocytes' nuclei segmentation. The proposed technique is being tested on a set of 365 images and the segmentation results are validated both qualitatively and quantitatively with current state-of-the-art methods on the basis of ground truth data (manually marked images by medical experts). The proposed technique is qualitatively compared with current state-of-the-art methods on the basis of ground truth data through visual inspection on four different grounds. Finally, the proposed technique quantitatively achieved an overall segmentation accuracy, sensitivity and precision of 92.8%, 93.5% and 98.16% respectively while an overall F-measure of 95.75%.

Removal of pectoral muscle based on topographic map and shape-shifting silhouette.

BACKGROUND: In digital mammography, finding accurate breast profile segmentation of women's mammogram is considered a challenging task. The existence of the pectoral muscle may mislead the diagnosis of cancer due to its high-level similarity to breast body. In addition, some other challenges due to manifestation of the breast body pectoral muscle in the mammogram data include inaccurate estimation of the density level and assessment of the cancer cell. The discrete differentiation operator has been proven to eliminate the pectoral muscle before the analysis processing. METHODS: We propose a novel approach to remove the pectoral muscle in terms of the mediolateral-oblique observation of a mammogram using a discrete differentiation operator. This is used to detect the edges boundaries and to approximate the gradient value of the intensity function. Further refinement is achieved using a convex hull technique. This method is implemented on dataset provided by MIAS and 20 contrast enhanced digital mammographic images. RESULTS: To assess the performance of the proposed method, visual inspections by radiologist as well as calculation based on well-known metrics are observed. For calculation of performance metrics, the given pixels in pectoral muscle region of the input scans are calculated as ground truth. CONCLUSIONS: Our approach tolerates an extensive variety of the pectoral muscle geometries with minimum risk of bias in breast profile than existing techniques.

Report: Nuclei segmentation of leukocytes in blood smear digital images.

The Leukocytes are differentiated from each other on the basis of their nuclei, demanded in many Medical studies, especially in all types of Leukemia by the Hematologists to note the disorder caused by specific type of Leukocyte. Leukemia is a life threatening disease. The work for diagnosing is manually carried out by the Hematologists involving much labor, time and human errors. The problems mentioned are easily addressed through computer vision techniques, but still accuracy and efficiency are demanded in terms of the basic and challenging step segmentation of Leukocyte's nuclei. The underlying study proposed better method in terms of accuracy and efficiency by designing a dynamic convolution filter for boosting low intensity values in the separated green channel of an RGB image and suppressing the high values in the same channel. The high values in the green channel become 255 (background) while the nuclei always have low values in the green channel and thus clearly appear as foreground. The proposed technique is tested on 365 images achieving an overall accuracy of 95.89%, while improving the efficiency by 10%. The proposed technique achieved its targets in a realistic way by improving the accuracy as well as the efficiency and both are highly required in the area.