Search

Abstract Reactive oxygen species (ROS) serve vital physiological functions, but aberrant ROS production contributes to numerous diseases. Unfortunately, therapeutic progress targeting pathogenic ROS has been hindered by the limited understanding of whether the mechanisms driving pathogenic ROS differ from those governing physiological ROS generation. To address this knowledge gap, we utilised a cellular model of Parkinson’s disease (PD), as an exemplar of ROS-associated diseases. We exposed SH-SY5Y neuroblastoma cells to the PD-toxin, MPP+ (1-methyl-4-phenylpyridinium) and studied ROS upregulation leading to cell death, the primary cause of PD. We demonstrate: (1) MPP+ stimulates ROS production by raising cytoplasmic Ca2+ levels, rather than acting directly on mitochondria. (2) To raise the Ca2+, MPP+ co-stimulates NADPH oxidase-2 (NOX2) and the Transient Receptor Potential Melastatin2 (TRPM2) channel that form a positive feedback loop to support each other’s function. (3) Ca2+ exacerbates mitochondrial ROS (mtROS) production not directly, but via Zn2+. (4) Zn2+ promotes electron escape from respiratory complexes, predominantly from complex III, to generate mtROS. These conclusions are drawn from data, wherein inhibition of TRPM2 and NOX2, chelation of Ca2+ and Zn2+, and prevention of electron escape from complexes -all abolished the ability of MPP+ to induce mtROS production and the associated cell death. Furthermore, calcium ionophore mimicked the effects of MPP+, while Zn2+ ionophore replicated the effects of both MPP+ and Ca2+. Thus, we unveil a previously unrecognized signalling circuit involving NOX2, TRPM2, Ca2+, Zn2+, and complex III that drives cytotoxic ROS production. This circuit lies dormant in healthy cells but is triggered by pathogenic insults and could therefore represent a safe therapeutic target for PD and other ROS-linked diseases.

The rapid advancement of deep learning has generated considerable enthusiasm regarding its utilization in addressing medical imaging issues. Machine learning (ML) methods can help radiologists to diagnose breast cancer (BCs) barring invasive measures. Informative hand-crafted features are essential prerequisites for traditional machine learning classifiers to achieve accurate results, which are time-consuming to extract. In this paper, our deep learning algorithm is created to precisely identify breast cancers on screening mammograms, employing a training method that effectively utilizes training datasets with either full clinical annotation or solely the cancer status of the entire image. The proposed approach utilizes Lightweight Convolutional Neural Network (LWCNN) that allows automatic extraction features in an end-to-end manner. We have tested LWCNN model in two experiments. In the first experiment, the model was tested with two cases' original and enhancement datasets 1. It achieved 95 %, 93 %, 99 % and 98 % for training and testing accuracy respectively. In the second experiment, the model has been tested with two cases' original and enhancement datasets 2. It achieved 95 %, 91 %, 99 % and 92 % for training and testing accuracy respectively. Our proposed method, which uses various convolutional network to classify screening mammograms achieved exceptional performance when compared to other methods. The findings from these experiments clearly indicate that automatic deep learning techniques can be trained effectively to attain remarkable accuracy across a wide range of mammography datasets. This holds significant promise for improving clinical tools and reducing both false positive and false negative outcomes in screening mammography.

The process of identifying the optimal unknown variables for the creation of a precision fuel-cell performance forecasting model using optimization techniques is known as parameter identification of the proton exchange membrane fuel cell (PEMFC). Recognizing these factors is crucial for accurately forecasting and assessing the fuel cell's performance, as they may not always be included in the manufacturer's datasheet. Six optimization algorithms—the Walrus Optimizer (WO), the Tunicate Swarm Algorithm (TSA), the Harris Hawks Optimizer (HHO), the Heap Based Optimizer (HBO), the Chimp Optimization Algorithm (ChOA), and the Osprey Optimization Algorithm (OOA) were used to compute six unknown variables of a PEMFC. Also, the proposed WO method was compared with other published works' methods such as the Equilibrium Optimizer (EO), Manta Rays Foraging Optimizer (MRFO), Neural Network Algorithm (NNA), Artificial Ecosystem Optimizer (AEO), Slap Swarm Optimizer (SSO), and Vortex Search Approach with Differential Evolution (VSDE). Minimizing the sum squares error (SSE) between the estimated and measured cell voltages requires treating these six parameters as choice variables during optimization. The WO algorithm yielded an SSE of 1.945415603, followed by HBO, HHO, TSA, ChOA, and OOA. Given that WO accurately forecasted the fuel cell's performance, it is appropriate for the development of digital twins for fuel cell applications and control systems for the automobile industry. Furthermore, it was shown that the WO convergence speed was faster than the other approaches studied.

With the rapid progress of communication technology, the Internet of Things (IoT) has emerged as an essential element in our daily lives. Given that the IoT encompasses diverse devices that often have limited resources in terms of communication, computation, and storage. Consequently, the National Institute of Standards and Technology (NIST) has standardized several lightweight cryptographic algorithms for encryption and decryption, specifically designed to meet the needs of resource-constrained IoT devices. These cryptographic algorithms, known as authenticated encryption with associated data (AEAD), offer more than just confidentiality—they also guarantee information integrity and authentication. Unlike conventional encryption algorithms like AES, which solely provide confidentiality, AEAD algorithms encompass additional functionality to achieve authenticity. This eliminates the need for separate algorithms like message authentication codes to ensure authenticity. Therefore, by leveraging the characteristics of an AEAD protocol, it is possible to develop a lightweight authentication framework to mitigate the security risks inherent in public communication channels. Therefore, in this work, we designed the lightweight authentication protocol for the smart healthcare system (BLAP-SHS) using an AEAD mechanism. In order to do this, a session key must first be created for encrypted communication. This is done via a method called mutual authentication, which verifies the legitimacy of both the user and the server. The random-or-real methodology ensures the security of the derived session key, and the Scyther tool is used to assess BLAP-SHS’ resistance to man-in-the-middle and replay attacks. Through using the technique of informal security analysis, the resilience of BLAP-SHS against denial of service, and password-guessing threats are evaluated. By juxtaposing BLAP-SHS with other prominent authentication techniques, the usefulness of BLAP-SHS is also assessed in terms of computing and communication costs. We illustrate that the BLAP-SHS requires a reduction in computation cost ranging from [70.11% to 95.21%] and a reduction in communication resources ranging from [3.85% to 9.09%], as evidenced by our comparative study.

Deep learning has played a vital role in advancing medical research, particularly in brain tumor segmentation. Despite using numerous deep learning algorithms for this purpose, accurately and reliably segmenting brain tumors remains a significant challenge. Segmentation of precise tumors is essential for the effective treatment of brain diseases. While deep learning offers a range of algorithms for segmentation, they still face limitations when analyzing medical images due to the variations in tumor shape, size, and location. This study proposes a deep learning approach combining a Generative Adversarial Network (GAN) with transfer learning and auto-encoder techniques to enhance brain tumor segmentation. The GAN incorporates a generator and discriminator to generate superior segmentation outcomes. In the generator, we applied downsampling and upsampling for tumor segmentation. In addition, an auto-encoder is applied in which the encoder retains as much information as possible and then the decoder with those encodings reconstructs the image. The transfer learning technique is applied at the bottleneck using the DenseNet model. Combining auto-encoder techniques with transfer learning methodologies in GANs feature learning is enhanced, training time is reduced, and stability is increased. In this work, we enhanced the accuracy of brain tumor segmentation and even achieved better results for tumors having small sizes. We train and evaluate our proposed model using the publicly available BraTS 2021 dataset. The experimental result shows a dice score of 0.94 for the whole tumor, 0.86 for the tumor core, and 0.82 for the enhancing tumor. It is also shown that we achieve 2% to 4% higher accuracy than other methods.

There has been a significant amount of attention in recent years toward the utilization of artificial intelligence (AI) in the realm of legal decision-making. This growing pattern reveals a higher interest among academics and legal professionals in utilizing AI technologies to enhance a number of legal system components. Artificial intelligence (AI) tools, such as machine learning and natural language processing, possess the capacity to analyze vast quantities of legal data, extract valuable insights, and facilitate decision-making processes. The primary aim of this study is to develop a sophisticated framework for judicial decision-making that incorporates methodologies from artificial intelligence and utilizes the dataset from the European Court of Human Rights (ECHR). The utilization of this methodology holds promise in improving the decision-making procedures of legal professionals and reducing the laborious task of manually analyzing legal documents. As a result, this can lead to the facilitation of more accurate predictions of court rulings. Our research introduces a hybrid ensemble model designed specifically for smart court rulings. This innovative approach harnesses the benefits of pre-trained embeddings and large language models to accurately predict court decisions. By utilizing the power of pre-existing embeddings and incorporating the capabilities of advanced language models, our proposed model demonstrates enhanced predictive accuracy and efficiency in the context of court rulings. We also focus on the models’ feasible interpretability and highlight their ability to determine key factors in legal decision-making. We attain a notably high accuracy score of around 83%. Our research illuminates how large language models (LLMs) and advanced deep learning techniques can be utilized to predict legal outcomes.

The modelling and forecasting of personal conduct depend on the records that individuals contribute at some stage in the shared network. It is essential to conduct this analysis to forecast the pursuits, feelings, and personal options, ultimately resulting in a development in service efficiency. However, to get accurate predictions, it’s far essential to triumph over the mission of information segregation and retaining precision in evaluation. This paper offers the preference-based predictive behaviour analysis (PPBA) approach to address this problem. This approach aims to achieve the best possible accuracy stage in the desired identity. A study primarily based on analysis is done on mutated and segregated data in the proposed methodology, which uses transfer learning. A more sophisticated understanding of consumer options is made possible through the system of segregation, which is completed by studying desire deviations from the attitude of many inputs. Diverse mutations in deviation sites are recognized during the learning method, which aligns options for various statistics. State validations are completed primarily based on an individual’s preceding options and any novel deviations observed inside the present state of affairs. This involves figuring out novel deviations from preceding behaviours, which can be rooted in various person options, which, in the long run, results in the refinement of user conduct prediction and modelling across quite a few applications primarily based on social networks. The proposed PPBA achieves 14.82% high detection accuracy for different input values, 9.41% less analysis time, 7.49% less false positives, 9.5% less complexity, and 10.34% high preference ratio.

This paper presents a modified bonobo optimizer (MBO) that integrates the Gaussian local mutation, restart strategy, and random contraction strategy into the original bonobo optimizer (BO). BO, inspired by the unique reproductive schemes and fission-fusion social behaviors of bonobos, has previously demonstrated promising results in solving a range of optimization problems. With the new modifications, MBO seeks to improve exploration and exploitation abilities, achieving enhanced convergence speed and solution quality. The Gaussian local mutation aids in fine-tuning solutions by introducing localized variations, the restart strategy provides a mechanism to escape potential local optima, while the random contraction strategy ensures better global search capabilities. The enhanced MBO’s performance is critically assessed on the 10 and 100-dimensional CEC 2017 and 10 and 20-dimensional CEC 2022 benchmark suites, along with seven engineering optimization problems, including cantilever beam design, industrial refrigeration system design, welded beam design, speed reducer design, pressure vessel design, multi-product batch plant design, and three-bar truss design. The MBO algorithm exhibits significant improvements in optimization performance, evidenced by highly significant p-values (as low as 1.25E-11) in the Wilcoxon’s Signed Rank Test. Preliminary results indicate that the MBO exhibits a marked improvement in both solution accuracy and robustness over its predecessor and other state-of-the-art optimization algorithms such as original bonobo optimizer, sand cat swarm optimization, Chernobyl disaster optimizer, driving training-based optimization, Harris hawk optimizer, Archimedes optimization algorithm, smell agent optimizer, grasshopper optimization algorithm, particle swarm optimization, hybrid sine cosine algorithm with differential evolution, modified capuchin search algorithm, liver cancer algorithm, and modified chameleon swarm algorithm. The algorithm’s robust performance can be attributed to its accelerated convergence rate, stability across diverse functions, good exploration-exploitation behavior, and adaptability to high-dimensional and complex solution spaces. The systematic enhancement of proposed algorithm’s convergence capabilities positions it as a reliable and efficient tool for addressing challenging engineering optimization problems.

Abstract The rising risk of diabetes, particularly in emerging countries, highlights the importance of early detection. Manual prediction can be a challenging task, leading to the need for automatic approaches. The major challenge with biomedical datasets is data scarcity. Biomedical data is often difficult to obtain in large quantities, which can limit the ability to train deep learning models effectively. Biomedical data can be noisy and inconsistent, which can make it difficult to train accurate models. To overcome the above-mentioned challenges, this work presents a new framework for data modeling that is based on correlation measures between features and can be used to process data effectively for predicting diabetes. The standard, publicly available Pima Indians Medical Diabetes (PIMA) dataset is utilized to verify the effectiveness of the proposed techniques. Experiments using the PIMA dataset showed that the proposed data modeling method improved the accuracy of machine learning models by an average of 9%, with deep convolutional neural network models achieving an accuracy of 96.13%. Overall, this study demonstrates the effectiveness of the proposed strategy in the early and reliable prediction of diabetes.

The telecom sector is currently undergoing a digital transformation by integrating artificial intelligence (AI) and Internet of Things (IoT) technologies. Customer retention in this context relies on the application of autonomous AI methods for analyzing IoT device data patterns in relation to the offered service packages. One significant challenge in existing studies is treating churn recognition and customer segmentation as separate tasks, which diminishes overall system accuracy. This study introduces an innovative approach by leveraging a unified customer analytics platform that treats churn recognition and segmentation as a bi-level optimization problem. The proposed framework includes an Auto Machine Learning (AutoML) oversampling method, effectively handling three mixed datasets of customer churn features while addressing imbalanced-class distribution issues. To enhance performance, the study utilizes the strength of oversampling methods like synthetic minority oversampling technique for nominal and continuous features (SMOTE-NC) and synthetic minority oversampling with encoded nominal and continuous features (SMOTE-ENC). Performance evaluation, using 10-fold cross-validation, measures accuracy and F1-score. Simulation results demonstrate that the proposed strategy, particularly Random Forest (RF) with SMOTE-NC, outperforms standard methods with SMOTE. It achieves accuracy rates of 79.24%, 94.54%, and 69.57%, and F1-scores of 65.25%, 81.87%, and 45.62% for the IBM, Kaggle Telco and Cell2Cell datasets, respectively. The proposed method autonomously determines the number and density of clusters. Factor analysis employing Bayesian logistic regression identifies influential factors for accurate customer segmentation. Furthermore, the study segments consumers behaviorally and generates targeted recommendations for personalized service packages, benefiting decision-makers.

Quality of life is greatly affected by chronic wounds. It requires more intensive care than acute wounds. Schedule follow-up appointments with their doctor to track healing. Good wound treatment promotes healing and fewer problems. Wound care requires precise and reliable wound measurement to optimize patient treatment and outcomes according to evidence-based best practices. Images are used to objectively assess wound state by quantifying key healing parameters. Nevertheless, the robust segmentation of wound images is complex because of the high diversity of wound types and imaging conditions. This study proposes and evaluates a novel hybrid model developed for wound segmentation in medical images. The model combines advanced deep learning techniques with traditional image processing methods to improve the accuracy and reliability of wound segmentation. The main objective is to overcome the limitations of existing segmentation methods (UNet) by leveraging the combined advantages of both paradigms. In our investigation, we introduced a hybrid model architecture, wherein a ResNet34 is utilized as the encoder, and a UNet is employed as the decoder. The combination of ResNet34’s deep representation learning and UNet’s efficient feature extraction yields notable benefits. The architectural design successfully integrated high-level and low-level features, enabling the generation of segmentation maps with high precision and accuracy. Following the implementation of our model to the actual data, we were able to determine the following values for the Intersection over Union (IOU), Dice score, and accuracy: 0.973, 0.986, and 0.9736, respectively. According to the achieved results, the proposed method is more precise and accurate than the current state-of-the-art.

Text classification is a common task in natural language processing (NLP), where the objective is to assign predefined categories or labels to a given text. Detecting sarcasm and classifying sentiment and dialect in NLP has practical applications, including spam detection, topic classification, and sentiment analysis. However, sarcasm and sentimental expressions, such as irony, humor, or criticism, can be difficult to identify through traditional NLP methods due to their implicit nature. To address this, we propose a Modified Switch Transformer (MST) for detecting sarcasm and classifying sentiment and dialect in Arabic text data. Our approach includes two key contributions: Variational Enmesh Expert’s Routing ( $VE_{e}R$ ) and Probabilistic Projections ( $P_{\phi} $ ). The switch transformer model incorporates probabilistic projections using a Variational Spatial Gated Unit-MLP to enhance the embedding generation mechanism. This updated mechanism introduces a variational aspect, providing dynamic control over the flow of information in the network, in contrast to the simpler embedding generation phase used in the original switch transformer. Moreover, we incorporate Variational Enmesh Expert’s Routing, which utilizes a hierarchical set of Variational experts, where each expert is a small and variational-directed acyclic graph network. The $VE_{e}R$ routing technique allows the network to dynamically choose which path to take at each layer based on the input, using a set of weights learned during training to determine the best route for a given input. Instead of optimizing route paths deterministically, we utilize Variational Inference and model each route as a random variable from a distribution. Our study evaluates the effectiveness of the Modified Switch Transformer (MST) model on the ArSarcasm Dataset, which includes Arabic language data related to sarcasm, dialect, and sentiments. We compare the performance of our proposed model with existing state-of-the-art models in the literature. The results show that the switch transformer outperforms other models in detecting sarcasm and also performs well in classifying sentiment and dialect.

ICU readmission is usually associated with an increased number of hospital death. Predicting readmission helps to reduce such risks by avoiding early discharge, providing appropriate intervention, and planning for patient placement after ICU discharge. Unfortunately, ICU scores such as the simplified acute physiology score (SAPS) and Acute Physiology and Chronic Health (APACHE) could help predict mortality or evaluate illness severity. Still, it is ineffective in predicting ICU readmission. This study introduces a clinical monitoring fog-computing-based system for remote prognosis and monitoring of intensive care patients. This proposed monitoring system uses the advantages of machine learning (ML) approaches for generating a real-time alert signal to doctors for supplying e-healthcare, accelerating decision-making, and monitoring and controlling health systems. The proposed system includes three main layers. First, the data acquisition layer, in which we collect the vital signs and lab tests of the patient’s health conditions in real-time. Then, the fog computing layer processes. The results are then sent to the cloud layer, which offers sizable storage space for patient healthcare. Demographic data, lab tests, and vital signs are aggregated from the MIMIC III dataset for 10,465 patients. Feature selection methods: Genetic algorithm (GA) and practical swarm optimization (PSO) are used to choose the optimal feature subset from detests. Moreover, Different traditional ML models, ensemble learning models, and the proposed stacking models are applied to full features and selected features to predict readmission after 30 days of ICU discharge. The proposed stacking models recorded the highest performance compared to other models. The proposed stacking ensemble model with selected features by POS achieved promising results (accuracy = 98.42, precision = 98.42, recall = 98.42, and F1-Score = 98.42), compared to full features and selected features. We also, provide model explanations to ensure efficiency, effectiveness, and trust in the developed model through local and global explanations.

The quality of software systems may be seriously impacted by specific types of source code anomalies. For example, poor programming practices result in Code Smells (CSs), which are a specific type of source code anomalies. They lead to architectural problems that consequently impact some significant software quality attributes, such as maintainability, portability, and reuse. To reduce the risk of introducing CSs and alleviate their consequences, the knowledge and skills of developers and architects is essential. On the other hand, ontologies, which are an artificial intelligence technique, have been used as a solution to deal with different software engineering challenges. Hence, the aim of this paper is to describe an ontological approach to representing and analyzing code smells. Since ontologies are a formal language based on description logics, this approach may contribute to formally analyzing the information about code smells, for example, to detect inconsistencies or infer new knowledge with the support of a reasoner. In addition, this proposal may support the training of software developers by providing the most relevant information on code smells. This ontology can also be a means of representing the knowledge on CSs from different sources (documents in natural language, relational databases, HTML documents, etc.). Therefore, it could be a valuable knowledge base to support the struggle of software developers and architects either to avoid CSs or to detect and remove them. The ontology was developed following a sound methodology. The well-known tool Protégé was used to manage the ontology and it was validated by using different techniques. An experiment was conducted to demonstrate the applicability of the ontology and evaluate its impact on speeding up the analysis of CSs.

Human pose and gesture estimation are crucial in correcting physiotherapy fitness exercises. In recent years, advancements in computer vision and machine learning approaches have led to the development of sophisticated pose estimation models that accurately track and analyze human movements in real time. This technology enables physiotherapists and fitness trainers to gain valuable insights into their client’s exercise forms and techniques, facilitating more effective exercise corrections and personalized training regimens. This research aims to propose an efficient artificial intelligence method for human pose estimation during physiotherapy fitness exercises. We utilized a multi-class exercise dataset based on human skeleton movement points to conduct our experimental research. The dataset comprises 133 features derived from human skeleton movements during various exercises, resulting in high feature dimensionality that affects the performance of human pose estimation with machine learning and deep learning methods. We have introduced a novel Logistic regression Recursive Feature elimination (LogRF) method for feature selection. Extensive experiments demonstrate that using the top twenty selected features, the random forest method outperformed state-of-the-art studies with a high-performance score of 0.998. The performance of each applied method is validated through a k-fold approach and further enhanced using hyperparameter tuning. Our proposed study assists specialists in identifying and addressing potential biomechanical issues, improper postures, and incorrect movement patterns, which are essential for injury prevention and optimizing exercise outcomes. Furthermore, this study enhances the capabilities of remote monitoring and guidance capabilities, allowing physiotherapists to support their patient’s progress with prescribed exercises continually.

The literature has proved the high efficiency of activated carbon in removing a wide range of pollutants from water; however, the high cost of industrial types of activated carbon is one of the barriers to the wide application of this efficient adsorbent. Therefore, this study aims at the development of a new type of activated carbon from the Schanginia/sp plant (Schanginia activated carbon (SAC)) and the possibility of using it as a biosorbent for the removal of methylene blue dye (MBD) from water. The SAC was subjected to characterisation analysis, including the surface area (SBET), Fourier-transform infrared spectroscopy (FTIR), scanning electron microscope (SEM), and statistical parametric mapping (SPM) before using it to remove the MBD. Furthermore, the efficacy of SAC was investigated in batch experiments, taking into account the effects of several experimental factors, namely pH of the solution, contact duration, SAC dose, and MBD concentration. The results showed the maximum uptake of MBD was 33.34 mg/g, and the Langmuir model is suitable to reproduce the optimum biosorption. In conclusion, the obtained results proved the applicability of SAC for the removal of MBD from water. HIGHLIGHTS A new type of activated carbon was made from natural sources.; The effects of operating parameters was studied.; Characterisation of the adsorbent was conducted using SBET, FTIR, SEM, and SPM.; The new adsorbent removed 33.34 mg/g of dyes in water.;

Light management can optimize algal biomass and biochemical composition, but responses vary based on growth stage and harvesting time. A 22 factorial design along with 3 center point runs was used to investigate the effects of light intensity (50, 100, and 150 μmol photon m−2 s−1) and exposure time (10.5, 16, and 21.5 h) on lipid, protein, carbohydrate, and pigments content, density, and productivity of Spirulina platensis. By analyzing trends in cellular content, the time each key metabolite reached its maximum was chosen as its best harvesting time. Low factor levels during the mid-exponential phase resulted in the highest protein content, while high levels at the very early exponential phase showed the most carbohydrates. Low light intensity favored the accumulation of lipids, chlorophyll, and carotenoids at the end of exponential phase. The highest phycobiliproteins were obtained at intermediate factor levels during the late exponential phase. Harvesting at high levels of factors at the end of exponential phase led to the highest component density. The maximum productivities were observed at high factor levels. These findings emphasize the importance of carefully selecting the harvesting time in different light conditions to achieve the desired biochemical content, density, and productivity in Spirulina applications.

Background: The study aimed to identify the types of bacteria isolated from the urine of patients with urinary tract infections focusing on the resistant isolates then comparing the results with the types of antibiotics misused by those patients in order to suggest some measures to mitigate the increasing rates of antibiotic resistance. Material and Methods: Urine samples were collected from patients aged from 10 to 60 years old who visited Al-Jomhory Hospital in Mosul City/Nineveh Governorate during a period between February and May 2021. Bacterial culture, identification, and antimicrobial susceptibility testing against different antibiotics were performed by Kirby Bauer's disc diffusion method and the results were compared with patient medical history of antibiotic use without consulting a physician. Results: Overall, 7 different bacterial pathogens were identified, 5 gram-negative and 2 gram-positive bacteria. The majority of bacterial pathogens isolated was Escherichia coli 37.6%, followed by Klebsiella spp. 35.5%. It is worth noting that the most effective drug was quinolone (P≤ 0.05) on bacterial species with least effective was penicillins worrisome results was the isolation of a high percentage of Pseudomonas aeruginosa, that was resistant to all antibiotics used in this research, by contrast, a high percentage of Enterococcus faecalis was sensitive to those antibiotics. Conclusion: The study revealed that a higher percentage of resistant bacteria were isolated in patients with a medical history of antibiotics misuse which might change drug prescribtion line in a hospitalized patients with bacterial infections.

The year 2020 brought many changes to the lives of people all over the world with the outbreak of COVID-19; we saw lockdowns for months and deaths of many individuals, which set the world economy back miles. As research was conducted to create vaccines and cures that would eradicate the virus, precautionary measures were imposed on people to help reduce the spread the disease. These measures included washing of hands, appropriate distancing in social gatherings and wearing of masks to cover the face and nose. But due to human error, most people failed to adhere to this face mask rule and this could be monitored using artificial intelligence. In this work, we carried out a survey on Masked Face Recognition (MFR) and Occluded Face Recognition (OFR) deep learning techniques used to detect whether a face mask was being worn. The major problem faced by these models is that people often wear face masks incorrectly, either not covering the nose or mouth, which is equivalent to not wearing it at all. The deep learning algorithms detected the covered features on the face to ensure that the correct parts of the face were covered and had amazingly effective results.

Identifying Human Epithelial Type 2 (HEp-2) mitotic cells is a crucial procedure in anti-nuclear antibodies (ANAs) testing, which is the standard protocol for detecting connective tissue diseases (CTD). Due to the low throughput and labor-subjectivity of the ANAs’ manual screening test, there is a need to develop a reliable HEp-2 computer-aided diagnosis (CAD) system. The automatic detection of mitotic cells from the microscopic HEp-2 specimen images is an essential step to support the diagnosis process and enhance the throughput of this test. This work proposes a deep active learning (DAL) approach to overcoming the cell labeling challenge. Moreover, deep learning detectors are tailored to automatically identify the mitotic cells directly in the entire microscopic HEp-2 specimen images, avoiding the segmentation step. The proposed framework is validated using the I3A Task-2 dataset over 5-fold cross-validation trials. Using the YOLO predictor, promising mitotic cell prediction results are achieved with an average of 90.011% recall, 88.307% precision, and 81.531% mAP. Whereas, average scores of 86.986% recall, 85.282% precision, and 78.506% mAP are obtained using the Faster R-CNN predictor. Employing the DAL method over four labeling rounds effectively enhances the accuracy of the data annotation, and hence, improves the prediction performance. The proposed framework could be practically applicable to support medical personnel in making rapid and accurate decisions about the mitotic cells’ existence.

Melanoma, a kind of skin cancer that is very risky, is distinguished by uncontrolled cell multiplication. Melanoma detection is of the utmost significance in clinical practice because of the atypical border structure and the numerous types of tissue it can involve. The identification of melanoma is still a challenging process for color images, despite the fact that numerous approaches have been proposed in the research that has been done. In this research, we present a comprehensive system for the efficient and precise classification of skin lesions. The framework includes preprocessing, segmentation, feature extraction, and classification modules. Preprocessing with DullRazor eliminates skin-imaging hair artifacts. Next, Fully Connected Neural Network (FCNN) semantic segmentation extracts precise and obvious Regions of Interest (ROIs). We then extract relevant skin image features from ROIs using an enhanced Sobel Directional Pattern (SDP). For skin image analysis, Sobel Directional Pattern outperforms ABCD. Finally, a stacked Restricted Boltzmann Machine (RBM) classifies skin ROIs. Stacked RBMs accurately classify skin melanoma. The experiments have been conducted on five datasets: Pedro Hispano Hospital (PH2), International Skin Imaging Collaboration (ISIC 2016), ISIC 2017, Dermnet, and DermIS, and achieved an accuracy of 99.8%, 96.5%, 95.5%, 87.9%, and 97.6%, respectively. The results show that a stack of Restricted Boltzmann Machines is superior for categorizing skin cancer types using the proposed innovative SDP.

Alzheimer’s disease (AD) is a neurodegenerative disorder characterized by cognitive impairment and aberrant protein deposition in the brain. Therefore, the early detection of AD is crucial for the development of effective treatments and interventions, as the disease is more responsive to treatment in its early stages. It is worth mentioning that deep learning techniques have been successfully applied in recent years to a wide range of medical imaging tasks, including the detection of AD. These techniques have the ability to automatically learn and extract features from large datasets, making them well suited for the analysis of complex medical images. In this paper, we propose an improved lightweight deep learning model for the accurate detection of AD from magnetic resonance imaging (MRI) images. Our proposed model achieves high detection performance without the need for deeper layers and eliminates the use of traditional methods such as feature extraction and classification by combining them all into one stage. Furthermore, our proposed method consists of only seven layers, making the system less complex than other previous deep models and less time-consuming to process. We evaluate our proposed model using a publicly available Kaggle dataset, which contains a large number of records in a small dataset size of only 36 Megabytes. Our model achieved an overall accuracy of 99.22% for binary classification and 95.93% for multi-classification tasks, which outperformed other previous models. Our study is the first to combine all methods used in the publicly available Kaggle dataset for AD detection, enabling researchers to work on a dataset with new challenges. Our findings show the effectiveness of our lightweight deep learning framework to achieve high accuracy in the classification of AD.

Brain tumors (BTs) are deadly diseases that can strike people of every age, all over the world. Every year, thousands of people die of brain tumors. Brain-related diagnoses require caution, and even the smallest error in diagnosis can have negative repercussions. Medical errors in brain tumor diagnosis are common and frequently result in higher patient mortality rates. Magnetic resonance imaging (MRI) is widely used for tumor evaluation and detection. However, MRI generates large amounts of data, making manual segmentation difficult and laborious work, limiting the use of accurate measurements in clinical practice. As a result, automated and dependable segmentation methods are required. Automatic segmentation and early detection of brain tumors are difficult tasks in computer vision due to their high spatial and structural variability. Therefore, early diagnosis or detection and treatment are critical. Various traditional Machine learning (ML) techniques have been used to detect various types of brain tumors. The main issue with these models is that the features were manually extracted. To address the aforementioned insightful issues, this paper presents a hybrid deep transfer learning (GN-AlexNet) model of BT tri-classification (pituitary, meningioma, and glioma). The proposed model combines GoogleNet architecture with the AlexNet model by removing the five layers of GoogleNet and adding ten layers of the AlexNet model, which extracts features and classifies them automatically. On the same CE-MRI dataset, the proposed model was compared to transfer learning techniques (VGG-16, AlexNet, SqeezNet, ResNet, and MobileNet-V2) and ML/DL. The proposed model outperformed the current methods in terms of accuracy and sensitivity (accuracy of 99.51% and sensitivity of 98.90%).

BACKGROUND: Drug shortages are a serious and complex issue in any healthcare system. We conducted this study because the prevalence of drug shortages in Saudi Arabia is largely unknown, while there have been reports of shortages. OBJECTIVE: To explore the prevalence and characteristics of drug shortages as well as identify strategies to minimize their impact on patient care and safety in large hospitals. DESIGN: Questionnaire-based cross-sectional study. SETTING: Pharmacy departments in secondary and tertiary care hospitals in the city of Riyadh. SUBJECTS AND METHODS: Pharmacists in ten hospitals, categorized as Ministry of Health [MOH], MOH-affiliated medical cities, and non-MOH, were recruited using convenience sampling. The European Association of Hospital Pharmacists drug shortage questionnaire was administered to survey pharmacists about drug shortages in their hospitals. MAIN OUTCOME MEASURES: Percentages of drug class shortages, characteristics, and strategies to minimize impact on patient care and safety across each hospital sector. RESULTS: Of 200 pharmacists invited to participate, 120 pharmacists completed the questionnaire (60% response rate). Twenty-four percent were from MOH hospitals, 32% from MOH-affiliated medical cities, and 44% from non-MOH hospitals. A significantly higher percentage of pharmacists from MOH-affiliated medical cities (42.11%) reported encountering drug shortages on a daily basis compared to 13.79% and 15.09% of participants from MOH-hospitals and non-MOH hospitals, respectively (P=.001). The top three drug classes that ≥ 25% of participants reported having shortages of were cardiovascular, antineoplastic, and endocrine drugs. The two most common strategies that were reported to minimize the impact of drug shortages on patient care by more than 70% of participants were informing prescribers and recommending alternative drugs, and alerting hospital staff about the presence of drug shortages using new communication tools. CONCLUSIONS: The relatively high reported rates of drug shortages in some hospitals should encourage health policymakers to address this serious public health problem. LIMITATIONS: The generazibility of the study's findings were limited by the small sample size, convenience sampling technique, self-reported data, and the fact that only pharmacists were invited to participate.

The study examined the values system and its relationship to leadership behavior of sports-practicing and non-sports-practicing students in Al-Balqa Applied University, Jordan. The study used two tools to collect data, namely, values system and leadership behavior questionnaires. The study sample consisted of (217) students, (64) of whom are sports-practicing, and (153) are non-sports-practicing. The results showed that the degree of students' possession of the values system as a whole was high, in descending order, moral and democratic values, achievement and competition values, national values, physical skills values, social values, aesthetic values and finally financial gain values. Besides, it showed that students' leadership behavior degree was high, and there was a statistically significant difference at ([alpha] = 0.05) between the means of the sports-practicing, and non-sports-practicing students' estimates on values system items. Also, there was a statistically significant difference in their estimates on the leadership behavior scale in favor of sports-practicing students. The study recommends universities to develop sports activities and intensify training courses that teach students leadership behavior and link knowledge and skills to the values system in their events, programs, curriculum and field activities.