International Journal of Information Technology and Computer Science @ijitcs
Journal articles - International Journal of Information Technology and Computer Science
All articles: 1304
Scientific article
This study presents an integrated traffic monitoring system for accident detection, vehicle counting by type, and vehicle speed estimation using roadside Closed-Circuit Television (CCTV) footage and machine vision based on the YOLOv11 architecture. The proposed methodology comprises data collection from heterogeneous sources, data preprocessing and augmentation, model fine-tuning on a custom Vehicle–Accident dataset, system deployment through a web-based application, and real-world evaluation. The YOLOv11 models were optimized to detect multiple vehicle categories and clearly defined accident classes under real traffic conditions. Experimental results indicate that the YOLOv11 Large (l) model achieves superior detection performance, with 81.8% precision, 75.8% recall, 82.1% mAP50, and 53.3% mAP50–95. Real-world testing further confirms its effectiveness, yielding an object detection accuracy of 99.24% and low speed estimation errors, with Mean Absolute Percentage Error (MAPE) of 3.56% for video-based evaluation and 5.54% for real-time evaluation. In contrast, the YOLOv11 Nano (n) model offers faster inference and lower computational requirements but exhibits reduced robustness in complex accident scenarios. The trained models are deployed in an interactive web application supporting image, video, and real-time inputs, enabling practical traffic monitoring and decision support. Overall, the YOLOv11l-Vehicle-Accident model is identified as the most suitable configuration for accuracy-critical traffic management systems, while Nano variants are better suited for resource-constrained deployments.
Free
Accident Response Time Enhancement Using Drones: A Case Study in Najm for Insurance Services
Scientific article
One of the main reasons for mortality among people is traffic accidents. The percentage of traffic accidents in the world has increased to become the third in the expected causes of death in 2020. In Saudi Arabia, there are more than 460,000 car accidents every year. The number of car accidents in Saudi Arabia is rising, especially during busy periods such as Ramadan and the Hajj season. The Saudi Arabia’s government is making the required efforts to lower the nations of car accident rate. This paper suggests a business process improvement for car accident reports handled by Najm in accordance with the Saudi Vision 2030. According to drone success in many fields (e.g., entertainment, monitoring, and photography), the paper proposes using drones to respond to accident reports, which will help to expedite the process and minimize turnaround time. In addition, the drone provides quick accident response and recording scenes with accurate results. The Business Process Management (BPM) methodology is followed in this proposal. The model was validated by comparing before and after simulation results which shows a significant impact on performance about 40% regarding turnaround time. Therefore, using drones can enhance the process of accident response with Najm in Saudi Arabia.
Free
Accuracy evaluation of brain tumor detection using entropy-based image thresholding
Scientific article
In this paper, the accuracy of the entropy-based thresholding approaches in brain tumor detection framework is investigated. Entropies are information gain methods that have been used for image thresholding with various application and different image modalities. The accuracy of the existing entropies for image thresholding has been studied in general domain (e.g.: natural images) and were not compared thoroughly. Thus, a framework for brain tumor segmentation is proposed with the core process of the image thresholding, in order to evaluate the accuracy of the entropies. Five entropies, namely, Renyi, Maximum, Minimum, Tsallis and Kapur are evaluated. Moreover, the aggregation of entropies was implemented and evaluated. The results show that the maximum entropy is the best for brain tumor detection. Moreover, it was shown that aggregation of entropies output does not enhance the result, however, it works as automatic selection of the best result and produces the results with the highest accuracy.
Free
Accurate Anomaly Detection using Adaptive Monitoring and Fast Switching in SDN
Scientific article
Software defined networking (SDN) is rapidly evolving technology which provides a suitable environment for easily applying efficient monitoring policies on the networks. SDN provides a centralized control of the whole network from which monitoring of network traffic and resources can be done with ease. SDN promises to drastically simplify network monitoring and management and also enable rapid innovation of networks through network programmability. SDN architecture separates the control of the network from the forwarding devices. With the higher innovation provided by the SDN, security threats at open interfaces of SDN also increases significantly as an attacker can target the single centralized point i.e. controller, to attack the network. Hence, efficient adaptive monitoring and measurement is required to detect and prevent malicious activities inside the network. Various such techniques have already been proposed by many researchers. This paper describes a work of applying efficient adaptive monitoring on the network while maintaining the performance of the network considering monitoring overhead over the controller. This work represents effective bandwidth utilization for calculation of threshold range while applying anomaly detection rules for monitoring of the network. Accurate detection of anomalies is implemented and also allows valid users and applications to transfer the data without any restrictions inside the network which otherwise were considered as anomalies in previous technique due to fluctuation of data and narrow threshold window. The concept of fast switching also used to improve the processing speed and performance of the networks.
Free
Scientific article
The most popular way for people to share information is through social media. Several studies have been conducted using ML approaches like LSTM, SVM, BERT, GA, hybrid LSTM-SVM and Multi-View Attention Networks to recognize bogus news MVAN. Most traditional systems identify false news or true news exclusively, but discovering kind of false information and prioritizing false information is more difficult, and traditional algorithms offer poor textual classification accuracy. As a result, this study focuses on predicting COVID-19-related false information on Twitter along with prioritizing types of false information. The proposed lightweight recommendation-system consists of three phases such as preprocessing, feature extraction and classification. The preprocessing phase is performed to remove the unwanted data. After preprocessing, the BERT model is used to convert the word into binary vectors. Then these binary features are taken as the input of the classification phase. In this classification phase, a 4CL time distributed layer is introduced for effective feature selection to remove the detection burdens, and the Bi-GRU model is used in the classification phase. Proposed-method is implemented in Mat lab software and is carried out several performance-metrics, and there are three different datasets used for validating its performance. Proposed model's total accuracy is 97%, specificity is 98%, precision is 95%, and the error value is 0.02, demonstrating its effectiveness over current methods. The proposed social media research system can accurately predict false information, and recognized news may be offered to the user such that they can learn the truth about news on social media.
Free
Active Selection Constraints for Semi-supervised Clustering Algorithms
Scientific article
Semi.-supervised clustering algorithms aim to enhance the performance of clustering using the pairwise constraints. However, selecting these constraints randomly or improperly can minimize the performance of clustering in certain situations and with different applications. In this paper, we select the most informative constraints to improve semi-supervised clustering algorithms. We present an active selection of constraints, including active must.-link (AML) and active cannot.-link (ACL) constraints. Based on Radial-Bases Function, we compute lower-bound and upper-bound between data points to select the constraints that improve the performance. We test the proposed algorithm with the base-line methods and show that our proposed active pairwise constraints outperform other algorithms.
Free
Scientific article
An evolving weighted neuro-neo-fuzzy-ANARX model and its learning procedures are introduced in the article. This system is basically used for time series forecasting. It's based on neo-fuzzy elements. This system may be considered as a pool of elements that process data in a parallel manner. The proposed evolving system may provide online processing data streams.
Free
Adaptive Guidance based on Context Profile for Software Process Modeling
Scientific article
This paper aims to define an adaptive guidance for software process modeling. The proposed guidance approach is based on development’s profile context (actor’s role in the process, actor’s qualification and related activities in progress). We introduce new guidance concepts through adaptive guidance meta-model (AGM) allowing specific assistance interventions (corrective, constructive and automatic guidance). We illustrate our guidance approach using SPEM formalism extended with these new guidance concepts.
Free
Scientific article
Polarimetric radar images suffer from the presence of speckles that degrade the received signal and introduce untruthful indications about the nature of the objects. In this study, we proposed a new framework to filter polarimetric images in which the edges and the channel correlation are preserved. Through a proposed scheme, the image is segmented into groups of regular and irregular pixels. The segmentation process is based on the homogeneity of the texture variation throughout the image. In the homogeneous area, speckle reduction is performed using the adaptive local mean of the neighboring pixels. For non-homogeneous surfaces, the scheme works independently for each set of resolution cells using the general product model containing both intensity and texture information. Quantitative and qualitative assessments confirmed that the proposed filter achieved highly ranked order; it has the ability to preserve fine details, polarimetric information, and to maintain the scattering mechanism of the different objects.
Free
Adaptive Modeling of Urban Dynamics during Armada Event using CDRs
Scientific article
This study investigates the mobile phone data during ephemeral event (Armada). The statistical techniques have been used for modeling human mobility collectively and individually. The undertaken substantial parameters are: inter-event times, travel distances (displacements), and radius of gyration. They have been analyzed and simulated using computing platform by integrating various applications for huge database management, visualization, analysis, and simulation. Accordingly, the general population pattern law has been extracted. This study has revealed the individuals mobility in dynamic perspective for 615,712 mobile users, also the simulated observed data are classified according to general, work, and off days.
Free
Adaptive Swarm-Optimized Ensemble Learning for Generalizable Heart Disease Risk Prediction
Scientific article
Machine learning (ML) has made it much easier to find and estimate the risk of early stage of cardiovascular illnesses by making it possible to analyses massive, various clinical datasets quickly and easily. In these kinds of datasets, demographic information, lifestyle characteristics, medical history, and diagnostic measurements are all included. These are all things that may not be easy to see through standard clinical examination. This study examines heart disease prediction through a series of hybrid ML models that integrate neighborhood-based classifiers, swarm intelligence-driven optimization, and ensemble learning, motivated by existing obstacles. There are four hybrid models being proposed: MSMO-KE and MSMO-KM, which combine Modified Spider Monkey Optimization (MSMO) with K-Nearest Neighbour classifiers that use Euclidean and Minkowski distance measures, respectively. There are also two ensemble variants, MSMO-KECB and MSMO-KMCB, which add CatBoost as a final prediction layer. To make sure it is strong and can be used in other situations, the proposed framework is tested on three separate cardiovascular datasets using a cross-validation method. The experimental findings show that the performance is always better than the baseline and the best models that are already used. The MSMO-KMCB model performs the best overall out of all the approaches tested. It has a cross-validated accuracy of 98.2% on Dataset-3 while keeping a high sensitivity. The comparative research demonstrates that the proposed MSMO-based ensemble models surpass current methodologies in predictive accuracy and recall, underscoring their promise for dependable and efficient heart disease risk prediction in clinical decision-support systems.
Free
Scientific article
The world is going to be a universal digital village and from the flow of this digitalization Bangladesh also riding of the tide. The key points of these digitalization is young generations basically university students of Bangladesh. ICT and Internet is a new trend for this country that’s the main reasons to encompass this by opportunity by young students. Bangladeshi young generations have also addicted in the upper tier of this list. The addiction of ICT & internet is more on the young generations than any other parts of the people generally in the third world countries. The main objective of this paper is to investigate the excessive use of Information and Communication Technology (ICT) and internet by the university students in Bangladesh. The study had collected the data from 24 public and private universities in Bangladesh out of 135. The study was used the simple random sampling (SRS) for analyzing the sample size with IBM SPSS 23.
Free
Advance Mining of Temporal High Utility Itemset
Scientific article
The stock market domain is a dynamic and unpredictable environment. Traditional techniques, such as fundamental and technical analysis can provide investors with some tools for managing their stocks and predicting their prices. However, these techniques cannot discover all the possible relations between stocks and thus there is a need for a different approach that will provide a deeper kind of analysis. Data mining can be used extensively in the financial markets and help in stock-price forecasting. Therefore, we propose in this paper a portfolio management solution with business intelligence characteristics. We know that the temporal high utility itemsets are the itemsets with support larger than a pre-specified threshold in current time window of data stream. Discovery of temporal high utility itemsets is an important process for mining interesting patterns like association rules from data streams. We proposed the novel algorithm for temporal association mining with utility approach. This make us to find the temporal high utility itemset which can generate less candidate itemsets.
Free
Advanced Applications of Neural Networks and Artificial Intelligence: A Review
Scientific article
Artificial Neural Network is a branch of Artificial intelligence and has been accepted as a new computing technology in computer science fields. This paper reviews the field of Artificial intelligence and focusing on recent applications which uses Artificial Neural Networks (ANN’s) and Artificial Intelligence (AI). It also considers the integration of neural networks with other computing methods Such as fuzzy logic to enhance the interpretation ability of data. Artificial Neural Networks is considers as major soft-computing technology and have been extensively studied and applied during the last two decades. The most general applications where neural networks are most widely used for problem solving are in pattern recognition, data analysis, control and clustering. Artificial Neural Networks have abundant features including high processing speeds and the ability to learn the solution to a problem from a set of examples. The main aim of this paper is to explore the recent applications of Neural Networks and Artificial Intelligence and provides an overview of the field, where the AI & ANN’s are used and discusses the critical role of AI & NN played in different areas.
Free
Advanced Deep Learning Models for Accurate Retinal Disease State Detection
Scientific article
Retinal diseases are a significant challenge in the realm of medical diagnosis, with potential complications to vision and overall ocular health. This research endeavors to address the challenge of automating the detection of retinal disease states using advanced deep learning models, including VGG-19, ResNet-50, InceptionV3, and EfficientNetV2. Each model leverages transfer learning, drawing insights from a substantial dataset comprising optical coherence tomography (OCT) images and subsequently classifying images into four distinct retinal conditions: choroidal neovascularization, drusen, diabetic macular edema and a healthy state. The training dataset, sourced from repositories that are available to the public including OCT retinal images, spanning all four disease categories. Our findings reveal that among the models tested, EfficientNetV2 demonstrates superior performance, with a remarkable classification accuracy of 98.92%, precision of 99.6%, and a recall of 99.4%, surpassing the performance of the other models.
Free
Scientific article
Text document clustering plays a pivotal role in organizing large-scale unstructured data, yet conventional clustering algorithms, such as k-means, often struggle with high-dimensional data, suboptimal initializations, and local minima issues. This paper introduces a novel comparative analysis of three advanced optimization techniques integrated with k-means: Chaotic Northern Goshawk Optimization (CNGO), Mayfly Optimization Algorithm (MOA), and Modified Particle Swarm Optimization (MPSO). This work is unique because it integrates these metaheuristic algorithms to improve clustering performance, targeting initialization challenges and increasing accuracy. Extensive experiments were conducted on benchmark datasets, including Reuters-21578, 20-Newsgroup, and BBC-Sport. All three models outper- form traditional k-means in terms of accuracy, Adjusted Rand Index (ARI), Normalized Mutual Information (NMI), and V-measure.This research offers new insights into optimizing clustering processes using metaheuristic algorithms and provides a foundation for future exploration in large-scale document clustering. Our study is significant because it systematically overcomes key limitations of conventional k-means for high-dimensional text data poor centroid initialization, local minima, and reduced effectiveness on sparse corpora by integrating and comparatively evaluating three advanced metaheuristics (CNGO, MOA, MPSO) with k-means on standard benchmark datasets. The value of this work lies in the consistently improved clustering quality (Accuracy, ARI, NMI, V-measure) achieved by the proposed hybrids, and in showing that MOA–k-means in particular offers a robust, scalable solution for real-world text analytics applications such as information retrieval, recommendation, and topic discovery.
Free
Advanced Prediction Based Mobility Support for 6LoWPAN Wireless Sensor Networks
Scientific article
Wireless Sensor Nodes (SNs), the key elements for building Internet of Things (IOT) have been deployed widely in order to get and transmit information over the internet. IPv6 over low power personal area network (6LoWPAN) enabled their connectivity with IPV6 networks. 6LoWPAN has mobility and it can find an extensive application space only if provides mobility support efficiently. Existing mobility schemes are focused on reducing handoff (HO) latency and pay less attention towards packet loss and signaling cost. In time critical applications under IOT, packet loss and excessive signaling cost are not acceptable. This paper proposes a scheme based on advanced mobility prediction for reducing extra signaling cost and packet loss that incurs due to connection termination in traditional schemes such as Proxy Mobile IPv6 (PMIPv6) handover. In our proposed scheme 6LoWPAN WSN architecture with IPv6 addressing is presented. Based on this architecture the mobility algorithm is proposed for reducing signaling cost, packet loss by buffering mechanism and HO latency in particular. In the algorithm layer 2 (L2) and layer 3 (L3) HO is performed simultaneously with prior HO prediction with no Care of Address (CoA) configuration which also reduces signaling cost to some extent. The proposed scheme is analyzed theoretically and evaluated for different performance metrics. Data results showed significant improvements in reducing packet loss, signaling cost and HO latency when compared to standard PMIPv6 in time critical scenarios.
Free
Scientific article
Network Intrusion Detection Systems (NIDS) play a vital role in modern cybersecurity by leveraging artificial intelligence (AI) in particular deep learning (DL) and machine learning (ML) to detect and mitigate malicious activities. However, these AI-driven systems are highly vulnerable to adversarial attacks, where small, imperceptible perturbations in input data can deceive models and significantly reduce detection accuracy. This raises critical concerns about the security and reliability of intrusion detection, especially in real-world scenarios where attackers exploit adversarial transferability to bypass defenses. This research investigates the threat posed by black-box adversarial attacks via surrogate models, focusing on the ability of adversarial examples to transfer across different architectures. This study simulates real-world adversarial threats, demonstrating how attacks crafted on one model can effectively deceive another, compromising NIDS security. A comparative study is conducted on two widely used AI models: an Artificial Neural Network (ANN) and a Convolutional Neural Network (CNN), both trained on the CICIDS 2019 dataset. The study evaluates the robustness of these models against two gradient-based adversarial attack methods, Fast Gradient Sign Method (FGSM) and Projected Gradient Descent (PGD), to determine their susceptibility under black box adversarial conditions. Experimental results indicate that CNN-based NIDS are more vulnerable to adversarial attacks than ANN-based models, with adversarial examples successfully transferring across architectures. These findings highlight the critical risks associated with adversarial transferability, underscoring the need for enhanced security measures to strengthen AI-driven intrusion detection systems against evolving cyber threats.
Free
Aerodynamic System Modeling based on Proper Orthogonal Decomposition
Scientific article
The main goal of present paper is to construct an efficient reduced order model (ROM) for aerodynamic system modeling. Proper Orthogonal Decomposition (POD) is presented to address the problem. First, the snapshots are collected to form the POD kernel, and then Singular Values Decomposition (SVD) is used to obtain POD modes, finally POD-ROM can be constructed by projecting full order aerodynamic system to POD modes subspace. Two problems are addressed: (1) aerodynamic data inverse design; (2) aeroelastic structure active control. For the second problem, POD method with balanced modification is introduced to improve the robustness of original POD method. Results in problem (1) suggest POD method works efficiently not only for interpolation inverse design but also for extrapolation problems. The results in problem (2) demonstrate POD method with balanced modification is efficient and accurate enough for aeroelastic system analysis.
Free
Airport Merchandising Using Micro Services Architecture
Scientific article
Airport offers an ecosystem where passengers, airlines, airport, and merchants meet under one roof during travel. During the journey, there is a good amount of time spent by passengers on waiting or transit before boarding the flight. Passengers look for spending that time by shopping, dining, and entertainment. Airport merchants and airlines look for merchandising opportunities to acquire new customers, up-sell, and cross-sell their products and services. Highly personalized merchandising offers can be generated on the fly by combining contextual information from passenger profile, likes and interests, merchant offers, and location specific events, seasons, and festivals. To achieve this, a strong airport merchandising platform is needed. The goal of the airport merchandising platform is to exchange information in a seamless manner across travel systems. The platform is designed on microservices design principles that use information from airlines, airport, social media, and merchant systems. Microservices can promote quick development, deployment, and realization of services. Microservices also improve the time to market capabilities. Mobile and desktop applications consume them to offer a personalized shopping experience to the passengers.
Free
