Search

Abstract Safety is crucial in the railway industry because railways transport millions of passengers and employees daily, making it paramount to prevent injuries and fatalities. In order to guarantee passenger safety, computer vision, unmanned aerial vehicles (UAV), and artificial intelligence will be essential tools in the near future for routinely evaluating the railway environment. An unmanned aerial vehicle captured dataset for railroad segmentation and obstacle detection (UAV-RSOD) comprises high-resolution images captured by UAVs over various obstacles within railroad scenes, enabling automatic railroad extraction and obstacle detection. The dataset includes 315 raw images, along with 630 labeled and 630 masked images for railroad semantic segmentation. The dataset consists of 315 original images captured by the UAV for object detection and obstacle detection. To increase dataset diversity for training purposes, we applied data augmentation techniques, which expanded the dataset to 2002 augmented and annotated images for obstacle detection cover six different classes of obstacles on railroad lines. Additionally, we provide the original 315 images along with a script for augmentation, allowing users to generate their own augmented data as needed, offering a more sustainable and customizable option. Each image in the dataset is accurately annotated with bounding boxes and labeled under six categories, including person, boulder, barrel, branch, jerry can, and iron rod. This comprehensive classification and detailed annotation make the dataset an essential tool for researchers and developers working on computer vision applications in the railroad domain.

Abstract This research presents a predictive model aimed at estimating the progression of Amyotrophic Lateral Sclerosis (ALS) based on clinical features collected from a dataset of 50 patients. Important features included evaluations of speech, mobility, and respiratory function. We utilized an XGBoost regression model to forecast scores on the ALS Functional Rating Scale (ALSFRS-R), achieving a training mean squared error (MSE) of 0.1651 and a testing MSE of 0.0073, with R² values of 0.9800 for training and 0.9993 for testing. The model demonstrates high accuracy, providing a useful tool for clinicians to track disease progression and enhance patient management and treatment strategies.

Abstract Insider threats pose a significant challenge in cybersecurity, demanding advanced detection methods for effective risk mitigation. This paper presents a comparative evaluation of data imbalance addressing techniques for CNN-based insider threat detection. Specifically, we integrate Convolutional Neural Networks (CNN) with three popular data imbalance addressing techniques: Synthetic Minority Over-sampling Technique (SMOTE), Borderline-SMOTE, and Adaptive Synthetic Sampling (ADASYN). The objective is to enhance insider threat detection accuracy and robustness in imbalanced datasets common to cybersecurity domains. Our study addresses the lack of consensus in the literature regarding the superiority of data imbalance addressing techniques in this field. We analyze a human behavior-based dataset (i.e., CERT) that reports users’ Information Technology (IT) activities with a substantial number of samples to provide a clear conclusion on the effectiveness of these balancing techniques when coupled with CNN. Experimental results demonstrate that ADASYN, in conjunction with CNN, achieves a ROC curve of 96%, surpassing SMOTE and Borderline-SMOTE in enhancing detection accuracy in imbalanced datasets. We compare the results of these three hybrid models (CNN + imbalance addressing techniques) with state-of-the-art selective studies focusing on ROC, recall, and accuracy measures. Our findings contribute to the advancement of insider threat detection methodologies.

Abstract The rise of Electric Vehicles (EVs) has introduced significant advancement and evolution in the electricity market. In smart transportation, the EVs have earned more popularity because of its numerous benefits including lower carbon footprints, higher performance, and sophisticated energy trading mechanisms. These potential benefits have resulted in widespread EV adoption across the world. Despite its benefits, energy management remains the biggest challenge in EVs and it is mainly because of the lack of Charging Stations (CSs) near EVs. This creates a demand for an effective, secure and reliable energy management framework for EVs. This study presents a secure data and energy trade paradigm based on Blockchain (BC) in the Internet of EVs (IoEV). BC technology prepares for the high volume of EV integration that serves as the foundation for the next generation, and to assist in developing unique privacy-protected BC-based D-Trading and storage Models. Entities evaluated for the proposed model include Trusted Authority (TA), Vehicles, Smart Meters, Roadside Units (RSU), BC, and Inter-Planetary File System (IPFS). In addition, E-trading involves several phases, including the acquiring E-trading demand requests, E-trading response requests, request matching and token assignment. Moreover, account mapping is performed using a Mayfly Pelican Optimization Algorithm (MPOA), which is created by merging the Mayfly Algorithm (MA) and Pelican Optimization Algorithm (POA). Various security features are used to protect data and energy trade in IoEV, including encryption, hashing, polynomials, and others. The testing results revealed that the MPOA outperformed the state-of-the-art results regarding memory consumption, trading rate, transaction cost, and trading energy volume with values of 4.605 MB, 91%, 0.654, and 90 kW, respectively.

Abstract The identification of anticancer peptides (ACPs) is crucial, especially in the development of peptide-based cancer therapy. The classical models such as Split Amino Acid Composition (SAAC) and Pseudo Amino Acid Composition (PseAAC) lack the incorporation of feature representation. These advancements improve the predictive accuracy and efficiency of ACP identification. Thus, the effort of this research is to propose and develop an advanced framework based on feature extraction. Thus, to achieve this objective herein we propose an Extended Dipeptide Composition (EDPC) framework. The proposed EDPC framework extends the dipeptide composition by considering the local sequence environment information and reforming the CD-HIT framework to remove noise and redundancy. To measure the accuracy, we have performed several experiments. These experiments were employed using four famous machine learning (ML) algorithms named; Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), and K Nearest Neighbor (KNN). For comparisons, we have used accuracy, specificity, sensitivity, precision, recall, and F1-Score as evaluation criteria. The reliability of the proposed framework is further evaluated using statistical significance tests. As a result, the proposed EDPC framework exhibited enhanced performance than SAAC and PseAAC, where the SVM model delivered the highest accuracy of 96. 6% and significant enhancements in specificity, sensitivity, precision, and F1-score over multiple datasets. Due to the incorporation of enhanced feature representation and the incorporation of local and global sequence profiles proposed EDPC achieves higher classification performance. The proposed frameworks can deal with noise and also duplicating features. These are accompanied by a wide range of feature representations. Finally, our proposed framework can be used for clinical applications where ACP identification is essential. Future works will include extending to a larger variety of datasets, incorporating tertiary structural information, and using deep learning techniques to improve the proposed EDPC.

Brain tumors are one of the leading causes of cancer death; screening early is the best strategy to diagnose and treat brain tumors. Magnetic Resonance Imaging (MRI) is extensively utilized for brain tumor diagnosis; nevertheless, achieving improved accuracy and performance, a critical challenge in most of the previously reported automated medical diagnostics, is a complex problem. The study introduces the Dual Vision Transformer-DSUNET model, which incorporates feature fusion techniques to provide precise and efficient differentiation between brain tumors and other brain regions by leveraging multi-modal MRI data. The impetus for this study arises from the necessity of automating the segmentation process of brain tumors in medical imaging, a critical component in the realms of diagnosis and therapy strategy. The BRATS 2020 dataset is employed to tackle this issue, an extensively utilized dataset for segmenting brain tumors. This dataset encompasses multi-modal MRI images, including T1-weighted, T2-weighted, T1Gd (contrast-enhanced), and FLAIR modalities. The proposed model incorporates the dual vision idea to comprehensively capture the heterogeneous properties of brain tumors across several imaging modalities. Moreover, feature fusion techniques are implemented to augment the amalgamation of data originating from several modalities, enhancing the accuracy and dependability of tumor segmentation. The Dual Vision Transformer-DSUNET model's performance is evaluated using the Dice Coefficient as a prevalent metric for quantifying segmentation accuracy. The results obtained from the experiment exhibit remarkable performance, with Dice Coefficient values of 91.47 % for enhanced tumors, 92.38 % for core tumors, and 90.88 % for edema. The cumulative Dice score for the entirety of the classes is 91.29 %. In addition, the model has a high level of accuracy, roughly 99.93 %, which underscores its durability and efficacy in segmenting brain tumors. Experimental findings demonstrate the integrity of the suggested architecture, which has quickly improved the detection accuracy of many brain diseases.

Breast cancer stands as one of the predominant health challenges globally, affecting millions of women every year and necessitating early and accurate detection to optimize patient outcomes. Currently, while deep convolutional neural networks (DCNNs) have shown promise in breast cancer detection, their application is often hampered by privacy concerns associated with sharing patient data and the limitation of training on small, localized datasets. Addressing these challenges, this manuscript introduces an effective federated learning approach tailored for breast cancer detection, leveraging DCNNs across diverse and large datasets without compromising data privacy. Our experimental findings underscore significant advancements in detection accuracy of 98.9% on three large scale datasets which are VINDR-MAMMO, CMMD, and INBREAST. Additionally, we tested the proposed federated learning performance, showcasing the potential of our approach as a robust and privacy-preserving solution for future breast cancer diagnostic strategies.

In today’s interconnected world, cybersecurity has emerged as a critical domain for ensuring the integrity, confidentiality, and availability of digital assets. Within this sphere, insider threats represent a unique and particularly insidious class of security risks, originating not from external hackers but from within the organization itself. These threats are perpetrated by individuals with inside information concerning the organization’s security practices, data, and computer systems. Traditional security measures like firewalls, intrusion detection systems, and antivirus software are often inadequate for tackling insider threats effectively, owing to their focus on external threats. This inadequacy underscores the urgent need for the development and implementation of more sophisticated, targeted detection techniques for insider threats. In response to this challenge, our research introduces an innovative approach that employs the Density-Based Local Outlier Factor (DBLOF) algorithm, fine-tuned to specifically tackle the challenges posed by the imbalanced nature of the CERT r4.2 insider threat dataset. This dataset is characterized by a highly skewed distribution, with a significant majority of benign instances and only a minimal proportion of malicious activities. Conventional detection algorithms often fail to effectively identify these rare but dangerous instances, leading to a high rate of false negatives. Our methodology capitalizes on the algorithm’s ability to focus on the local density deviation of data points, thereby enabling the precise identification of outliers that are indicative of potential insider threats. Through rigorous testing and validation processes, we have achieved outstanding results, with an of F-score 98%. These remarkable outcomes not only affirm the effectiveness of the DBLOF algorithm as a powerful tool for combating insider threats but also contribute valuable insights to the broader academic and professional discourse on cybersecurity. Importantly, our findings have practical implications, offering organizations actionable recommendations for boosting their internal security mechanisms against the complex and evolving landscape of insider threats.

Sentiment Analysis, a crucial tool for analyzing user opinions, has shown efficacy particularly when tailored to specific domains. While existing research predominantly focuses on training various classifiers for sentiment analysis within the software engineering (SE) domain, the outcomes often lack consistency when tested across different datasets. To address this gap, this paper proposes a novel approach utilizing transfer learning-based classifiers, fine-tuned and evaluated across diverse SE datasets. A comprehensive study is conducted, benchmarking machine learning and deep learning classifiers for SE sentiment analysis. Results indicate that transfer learning classifiers, namely GPT and BERT, outperform traditional approaches. Notably, the Bert large model achieves an F1-score of 0.89 on the Stack Overflow dataset, surpassing existing state-of-the-art tools. This research not only provides centralized insights but also paves the way for developing more accurate domain-specific sentiment analysis tools tailored for Software Engineering.

Diagnosing brain tumors using magnetic resonance imaging (MRI) presents significant challenges due to the complexities of segmentation and the variability in tumor characteristics. To address the limitations inherent in traditional methods, this research employs an advanced deep learning approach, integrating ResNet50 for feature extraction and Generative Adversarial Networks (GANs) for data augmentation. A comprehensive evaluation of ten transfer learning algorithms, including GoogLeNet and VGG-16, was conducted for the classification of brain tumors. Model performance was assessed using precision, recall, and F1-score metrics, complemented by additional metrics such as Hamming loss and the Matthews correlation coefficient to provide a more comprehensive insight. To ensure transparency in image predictions, Explainable AI techniques, specifically Local Interpretable Model-Agnostic Explanations (LIME), were utilized. The study involved the analysis of 7023 MRI images, with TumorGANet being trained on a dataset encompassing gliomas, meningiomas, non-tumorous cases, and pituitary tumors. The results demonstrate the exceptional performance of proposed model named TumorGANet, achieving an accuracy of 99.53%, precision and recall rates of 100%, F1 scores of 99%, and a Hamming loss of 0.2%.

Data owners can store extensive data in a flexible and cost-efficient manner by utilizing cloud data storage. Thus, the number of online cloud service providers and their users have undergone a substantial increase. A significant concern in the field of cloud computing is the security and confidentiality of data. A cryptographic access control technique can be used to guarantee data privacy, which prevents unauthorized access and allows only authorized users to access the data. Nevertheless, the trust issue is not fully addressed in this kind of cryptographic approach. The study primarily aims to identify a solution that effectively addresses the trust issues within access control methods, thereby reducing the security risks associated with cloud storage systems. This, in turn, enhances the quality of decisions made by data owners as well as cloud operators. In this paper, we proposed integrating task-role-based access control (T-RBAC) as a new cloud access control model with a comprehensive trust model using various criteria to offer high security for cloud storage systems, with taking flexibility into consideration. There are trust-based systems integrated with various existing access control models in cloud computing; however, these models do not provide accurate trust value computations, and may still be at risk of certain security threats. Hence, these threats can be eliminated through an accurate trust-based T-RBAC model.

The security of the Internet of Things (IoT) is crucial in various application platforms, such as the smart city monitoring system, which encompasses comprehensive monitoring of various conditions. Therefore, this study conducts an analysis on the utilization of blockchain technology for the purpose of monitoring Internet of Things (IoT) systems. The analysis is carried out by employing parametric objective functions. In the context of the Internet of Things (IoT), it is imperative to establish well-defined intervals for job execution, ensuring that the completion status of each action is promptly monitored and assessed. The major significance of proposed method is to integrate a blockchain technique with neuro-fuzzy algorithm thereby improving the security of data processing units in all smart city applications. As the entire process is carried out with IoT the security of data in both processing and storage units are not secured therefore confidence level of monitoring units are maximized at each state. Due to the integration process the proposed system model is implemented with minimum energy conservation where 93% of tasks are completed with improved security for about 90%.

Insider attacks may inflict far greater damage to an organization than outsider threats since insiders are authorized users who are acquainted with the business’s system, making detection harder. Many techniques to detecting insider threats have been developed, but they are neither flexible nor resilient owing to different obstacles (e.g., lack of real-world dataset and highly skewed class distribution of the available dataset), making insider threat detection an understudied research field. Previous techniques attempted to solve the dataset’s imbalance issue by increasing or lowering the observations of the dataset’s classes, however this might lead to underfitting and overfitting problems. We present an insider threat detection model that addresses the class imbalance problem at the algorithm level using anomaly-based techniques, as an enhancement over previous approaches. To limit the effect of skewed class distribution on insider threat detection, the Isolation Forest (IF) technique is used. The model is verified using the benchmarked CERT’s insider threat dataset, which is significantly unbalanced, with a small number of malicious cases vs a large number of non-malicious instances. Several contamination ratios of IF’s parameters are used to verify the model’s performance throughout a range of anomaly scores. The experimental findings reveal that the suggested model handles the dataset class imbalance problem with an accuracy score of 98%. The findings are compared to the baseline technique to demonstrate how the proposed model enhances detection performance and addresses the problem of data imbalance. The findings indicate the usefulness of the suggested approach for identifying insider threats when compared to previous studies.

Disabled people are facing many difficulties communicating with others and involving in society. Modern societies have dedicated significant efforts to promote the integration of disabled individuals into their societies and services. Currently, smart healthcare systems are used to facilitate disabled people. The objective of this paper is to enable two-way communication of deaf individuals with the rest of society, thus enabling their migration from marginal elements of society to mainstream contributing elements. In the proposed system, we developed three modules; the sign recognition module (SRM) that recognizes the signs of a deaf individual, the speech recognition and synthesis module (SRSM) that processes the speech of a non-deaf individual and converts it to text, and an Avatar module (AM) to generate and perform the corresponding sign of the non-deaf speech, which were integrated into the sign translation companion system called Saudi deaf companion system (SDCS) to facilitate the communication from the deaf to the hearing and vice versa. This paper also contributes to the literature by utilizing our self-developed database, the largest Saudi Sign Language (SSL) database—the King Saud University Saudi-SSL (KSU-SSL). The proposed SDCS system performs 293 Saudi signs that are recommended by the Saudi Association for Hearing Impairment (SAHI) from 10 domains (healthcare, common, alphabets, verbs, pronouns and adverbs, numbers, days, kings, family, and regions).

Recently, several countries have been trying hard to facilitate the integration of disabled people into their societies by ensuring equal opportunities through ease of access to social services, daily human necessities, and the labor market. Deafness is considered one of the major disabilities separating the deaf from their society. To integrate the deaf fully into society, a two-way mode of communication is required: one from the deaf to the hearing people, and the other from the hearing to the deaf. Communication from the hearing person to the deaf is generally easy and can be done through speech recognition and text-to-sign representations, but communication from the deaf to the hearing is somewhat difficult and requires a sign recognition module that recognizes the sign motions from the deaf and translates it to a text; following this, a speech synthesis module will translate this text to speech. To build the sign recognition module, a sign language dataset is required. This paper contributes to the literature by introducing a comprehensive survey of 17 Arabic sign language datasets and by developing a well-organized framework that is used to build a sign language dataset. This paper also contributes to the literature by creating the largest Saudi Sign Language (SSL) database—the King Saud University Saudi-SSL (KSU-SSL data-base)—with 293 signs, 33 signers, 145,035 samples, and 10 domains (healthcare, common, alpha-bets, verbs, pronouns and adverbs, numbers, days, kings, family, and regions). This paper also contributes to the literature by introducing a convolutional graph neural network (CGCN) architecture for sign language recognition and applying the proposed architecture to the built KSU-SSL database. The architecture is made up of a small number of separable 3DGCN layers, and is augmented with a spatial attention mechanism. This study is a part of the project that aims to develop a two-way communication system for Saudi deaf individuals.

Presents corrections to the paper, (Corrections to “Federated Learning Approach for Breast Cancer Detection Based on DCNN”).

Brain tumor segmentation from magnetic resonance imaging (MRI) scans is critical for the diagnosis, treatment planning, and monitoring of therapeutic outcomes. Thus, this research introduces a novel hybrid approach that combines handcrafted features with convolutional neural networks (CNNs) to enhance the performance of brain tumor segmentation. In this study, handcrafted features were extracted from MRI scans that included intensity-based, texture-based, and shape-based features. In parallel, a unique CNN architecture was developed and trained to detect the features from the data automatically. The proposed hybrid method was combined with the handcrafted features and the features identified by CNN in different pathways to a new CNN. In this study, the Brain Tumor Segmentation (BraTS) challenge dataset was used to measure the performance using a variety of assessment measures, for instance, segmentation accuracy, dice score, sensitivity, and specificity. The achieved results showed that our proposed approach outperformed the traditional handcrafted feature-based and individual CNN-based methods used for brain tumor segmentation. In addition, the incorporation of handcrafted features enhanced the performance of CNN, yielding a more robust and generalizable solution. This research has significant potential for real-world clinical applications where precise and efficient brain tumor segmentation is essential. Future research directions include investigating alternative feature fusion techniques and incorporating additional imaging modalities to further improve the proposed method’s performance.

Mobile edge computing (MEC) is a powerful new technology with the potential to transform and decentralize the way our cell phone networks currently work. The purpose of MEC is to process the intensive mobile applications in the available resources, which are embedded in the base station of the cell phone systems and closer to users (i.e., MEC support stations). We assumed that the telecommunications base station supports MEC, which provides edge computing with tiny latency. However, the problem of inevitable optimization emerges in terms of the quality of service (QoS) and user experience (QoE). Therefore, MEC services provide integrated services close to end-users to achieve QoS and QoE. This study examined how to jointly optimize resource allocation when offloading tasks from mobile devices (MD) to edge servers (ES) in MEC systems, thereby minimizing the computing time and service cost. The study’s main insight is that offloaded tasks can be delivered in a scheduled manner to the virtual machines (VMs) in the ES to minimize computing time, service cost, waste over the capability of the ES, and maximum associativity ( ${{\mathcal A}}_{E,X}$ ) of a task with an ES to maintain MD mobility. We present a dynamic task scheduling and load-balancing technique based on an integrated accelerated particle swarm optimization (APSO) algorithm with dynamic programming as a multi-objective. The proposed method was compared with the standard PSO, APSO, and PSO-GA algorithms using experimental simulations. The results show that the proposed method outperformed these algorithms, with a reduction in task makespan of 30% and an increase in resource utilization of 29% observed compared to PSO-GA. Additionally, the proposed method was associated with reducing service cost and waiting time compared to the other algorithms and improvements in the fitness function value.

In recent years, computation offloading has become an effective way to overcome the constraints of mobile devices (MDs) by offloading delay-sensitive and computation-intensive mobile application tasks to remote cloud-based data centers. Smart cities can benefit from offloading to edge points in the framework of the so-called cyber-physical-social systems (CPSS), as for example in traffic violation tracking cameras. We assume that there are mobile edge computing networks (MECNs) in more than one region, and they consist of multiple access points, multi-edge servers, and N MDs, where each MD has M independent real-time massive tasks. The MDs can connect to a MECN through the access points or the mobile network. Each task be can processed locally by the MD itself or remotely. There are three offloading options: nearest edge server, adjacent edge server, and remote cloud. We propose a reinforcement-learning-based state-action-reward-state-action (RL-SARSA) algorithm to resolve the resource management problem in the edge server, and make the optimal offloading decision for minimizing system cost, including energy consumption and computing time delay. We call this method OD-SARSA (offloading decision-based SARSA). We compared our proposed method with reinforcement learning based Q learning (RL-QL), and it is concluded that the performance of the former is superior to that of the latter.

Sign language is the main channel for hearing-impaired people to communicate with others. It is a visual language that conveys highly structured components of manual and non-manual parameters such that it needs a lot of effort to master by hearing people. Sign language recognition aims to facilitate this mastering difficulty and bridge the communication gap between hearing-impaired people and others. This study presents an efficient architecture for sign language recognition based on a convolutional graph neural network (GCN). The presented architecture consists of a few separable 3DGCN layers, which are enhanced by a spatial attention mechanism. The limited number of layers in the proposed architecture enables it to avoid the common over-smoothing problem in deep graph neural networks. Furthermore, the attention mechanism enhances the spatial context representation of the gestures. The proposed architecture is evaluated on different datasets and shows outstanding results.

The emerging technology of mobile cloud is introduced to overcome the constraints of mobile devices. We can achieve that by offloading resource intensive applications to remote cloud-based data centers. For the remote computing solution, mobile devices (MDs) experience higher response time and delay of the network, which negatively affects the real-time mobile user applications. In this study, we proposed a model to evaluate the efficiency of the close-end network computation offloading in MEC. This model helps in choosing the adjacent edge server from the surrounding edge servers. This helps to minimize the latency and increase the response time. To do so, we use a decision rule based Heuristic Virtual Value (HVV). The HVV is a mapping function based on the features of the edge server like the workload and performance. Furthermore, we propose availability of a virtual machine resource algorithm (AVM) based on the availability of VM in edge cloud servers for efficient resource allocation and task scheduling. The results of experiment simulation show that the proposed model can meet the response time requirements of different real-time services, improve the performance, and minimize the consumption of MD energy and the resource utilization.

In service-oriented computing, web services composition is the process of translating user requirements into a workflow. This workflow comprises many tasks, each of which includes an abstract definition for some of the user requirements. Web services can be aggregated to handle the workflow. Many of these services are available from various providers for each task; they are referred to, in aggregate, as the candidate list. The web service selection (WSS) problem centers on selecting the best service from these candidates based on the quality of service (QoS) features. In this paper, we propose an enhancement to the ant colony optimization (ACO) algorithm based on a swap concept for the QoS-aware WSS problem. The aim of the enhancement to the ACO is to avoid the trap of local optima and reduce the search duration. We believe that the integration of many potent solutions will help the ACO algorithm yield a better solution and avoid stagnation. Several experiments were conducted to compare the proposed algorithm with the ACO and flying ACO (FACO) algorithms. Two different types of experiments using 22 datasets were done with 30 independent repetitions. The first type of experiment's results shows that the proposed algorithm is better than ACO by 12% and FACO by 11% in terms of quality of solutions. The results in the second type of experiment show that the proposed algorithm continuously outperforms both algorithms in terms of quality of solutions.