Articles

Access the latest knowledge in applied science, electrical engineering, computer science and information technology, education, and health.

Advanced Search
Filter Icon

Filters article

Years

FAQ Arrow
Single Year
Range Year
0
0

Source Title

FAQ Arrow

Authors

FAQ Arrow

31,456 Article Results

Ensemble recursive feature elimination-based ensemble classification for medical diagnosis

10.11591/ijeecs.v40.i2.pp758-771
Thirumalaimuthu Thirumalaiappan Ramanathan , Md. Jakir Hossen , Abdullah Al Mamun , Joseph Emerson Raja
The application of data mining techniques for the extraction of patterns from medical datasets is useful in the prediction of various diseases from the data of patients. An appropriate feature selection method is required for the medical datasets to give better results for the medical data mining process. In data preprocessing, feature selection is an important process that finds the most relevant features from the dataset. Considering all features of the medical dataset without using any feature selection process may sometimes lead to inaccurate results. Most of the medical datasets contain meaningless data that are not relevant to the data mining process. These data can be eliminated through the feature selection process. This paper presents an integration of an ensemble feature selection approach and an ensemble classification approach through a classifier called the ensemble recursive feature elimination-based ensemble classifier (ERFE-EC) for the classification of medical data. Four different medical datasets were used for testing the ERFE-EC method, which showed promising results.
Volume: 40
Issue: 2
Page: 758-771
Publish at: 2025-11-01

FPGA-based implementation of an S-Box cryptographic co-processor for high-performance applications

10.11591/ijeecs.v40.i2.pp1167-1176
Moulai Khatir Ahmed Nassim , Ziani Zakarya
The increasing demand for reliable cryptographic operations for securing current systems has given birth to well-advanced and developed hardware solutions, in this paper we consider issues within the traditional symmetric advanced encryption standard (AES) cryptographic system as major challenges. Additionally, problems such as throughput limitations, reliability, and unified key management are also discussed and tackled through appropriate hierarchical transformation techniques. To overcome these challenges, this paper presents the design and field programmable gate array (FPGA)-based implementation of a cryptographic coprocessor optimized for substitution box (S-Box) operation which is considered as a key component in many cryptographic algorithms such as AES. The architecture of the co-processor proposed in this article is based on the advanced characteristics of FPGAs to accelerate the S-Box transformation, improve throughput and reduce latency compared to software implementations. We discussed carefully the design considerations along with resource utilization, speed optimization, and energy efficiency. The obtained experimental results present significant performance improvements, the FPGA-based implementation ensured higher throughput and lower execution time compared to traditional CPU-based methods. We presented in this work the effectiveness of using FPGAs for the acceleration of cryptographic operations in secure applications which will therefore be a robust solution for the next generation of secure systems.
Volume: 40
Issue: 2
Page: 1167-1176
Publish at: 2025-11-01

Image recognition using deep learning: a review

10.11591/ijeecs.v40.i2.pp953-967
Osama M. Hassan , Ashraf A. Gouda , Mohammed Abdel Razek
This paper presents a comprehensive review of recent advancements in image recognition, with a focus on deep learning (DL) techniques. Convolutional neural networks (CNNs), in particular, have significantly transformed this domain, enabling substantial improvements in both accuracy and efficiency across diverse applications. The review explores state-of-the-art methods, highlighting their practical implementations and the progress achieved. It also addresses key challenges such as data scarcity and model interpretability, offering perspectives on emerging opportunities and future directions. By synthesizing current trends with forward-looking insights, the paper aims to serve as a valuable resource for researchers and practitioners seeking to navigate and contribute to the evolving landscape of image recognition. Moreover, the paper examines critical challenges that persist in the field, such as transfer learning, data augmentation, and explainable artificial intelligence (AI) approaches. By synthesizing current trends with emerging innovations, the review not only maps the trajectory of progress but also highlights future directions and research opportunities. This synthesis aims to provide researchers, developers, and industry practitioners with a solid understanding of the dynamic and rapidly evolving environment surrounding image recognition technologies.
Volume: 40
Issue: 2
Page: 953-967
Publish at: 2025-11-01

Hyperparameter optimization of convolutional neural network using grey wolf optimization for facial emotion recognition

10.11591/ijeecs.v40.i2.pp898-906
Muhammad Munsarif , Muhammad Saman , Ernawati Ernawati , Budi Santosa
Facial emotion recognition (FER) is a challenging task in computer vision with wide applications in areas such as human-computer interaction, security, and healthcare. To improve the performance of convolutional neural networks (CNN) in FER, a novel approach combining CNN with grey wolf optimization (GWO) was proposed to optimize key hyperparameters. The CNN-GWO model was fine-tuned by adjusting hyperparameters such as the number of convolutional layers, kernel size, number of filters, and learning rate. This model was evaluated using the CK+ dataset and achieved an accuracy of 90.97%, demonstrating its competitive performance compared to existing methods. The optimized hyperparameters included three convolutional layers, 35 filters, a kernel size of 5, a learning rate of 0.045990, a dropout rate of 0.4988, and a max pooling size of 3. These results confirm that GWO is effective in optimizing CNN for FER tasks, providing an efficient solution to enhance model accuracy. This approach shows promising potential for future FER applications, highlighting GWO as a valuable optimization technique for CNN architectures.
Volume: 40
Issue: 2
Page: 898-906
Publish at: 2025-11-01

A novel approach for detection of cracks in painting and concrete surface images using CNN models

10.11591/ijeecs.v40.i2.pp988-1000
Deepti Vadicherla , Poonam Gupta
Discovering the beginnings of historical artworks takes one on an amazing voyage across space and time. People all around the world have been captivated by India's rich cultural heritage throughout its history, and ancient paintings have always been a very important part of it. Over the period of time, these ancient paintings can get cracks on it due to many factors. This research introduces an automated image classification system where the cracks on the paintings as well as the concrete surface will get detected. Detecting cracks on the concrete surface is important because the longevity and upkeep of concrete structures rely on the prompt identification and treatment of cracks, which can weaken the structure and necessitate expensive repairs. In this study, we focus on image classification using general convolution neural network (CNN), Inception V3, VGG-16, and ResNet-50 models of CNN. These models are trained and validated separately on two different datasets of paintings and concrete surfaces. Inception V3 and VGG-16 models achieve high accuracy, respectively in painting and concrete datasets in comparison with general CNN and ResNet-50 models.
Volume: 40
Issue: 2
Page: 988-1000
Publish at: 2025-11-01

Design and implementation of heterogeneous IoT wearables for multi-disease monitoring with OFDM-based spectrum allocation

10.11591/ijeecs.v40.i2.pp667-677
Shittu Moshood Boladale , Omotayo Olabowale Oshiga , Opeyemi Ayokunle Osanaiye , Abdulrasaq Olanrewaju Amuda , Abigail Chidimma Odigbo , Timothy Oluwaseun Araoye
This research proposes a comprehensive and scalable architecture for intelligent healthcare monitoring, integrating heterogeneous wearable biosensors, edge computing, and bio-inspired optimization techniques employing an orthogonal frequency division multiplexing (OFDM)-based spectrum allocation strategy. The system continuously monitors key physiological parameters, including heart rate, electrocardiogram (ECG), blood glucose levels, body temperature, blood pressure, and respiratory rate, using low-power, biocompatible sensors with wireless communication capabilities. An edge computing layer performs real-time signal preprocessing (noise filtering, normalization, compression), significantly reducing latency and bandwidth demands. To optimize system performance, the walrus optimization algorithm (WOA), a novel metaheuristic inspired by walrus social and hunting behaviors, is employed. WOA is utilized to dynamically adjust critical parameters, including transmission power, modulation index, bandwidth allocation, and routing efficiency. Experimental results demonstrate notable improvements: signal-to-noise ratio (SNR) increased from 5 dB to over 31 dB, latency reduced from 10 ms to under 4 ms, and bit error rate (BER) was minimized to 8×10⁻⁶. Hybrid models incorporating WOA with machine learning (WOA-ANN, WOA-SVM) achieved spectral efficiencies up to 3.7 bits/s/Hz and energy efficiencies up to 22 bits/Joule. The proposed system supports reliable, real-time health data acquisition and transmission in both urban and remote healthcare environments. Its modular, power-efficient, and adaptive architecture demonstrates high potential for deployment in telemedicine, chronic disease management, and emergency response systems, establishing a robust foundation for next-generation smart healthcare infrastructure.
Volume: 40
Issue: 2
Page: 667-677
Publish at: 2025-11-01

Monocular vision-based visual control for SCARA-type robotic arms: a depth mapping approach

10.11591/ijeecs.v40.i2.pp801-813
Diego Chambi Tula , Bryan Challco , Jonathan Catari , Walker Aguilar , Lizardo Pari
The accelerated growth of an increasingly automated industry requires the use of autonomous robotic systems. However, the use of these systems commonly requires an enormous amount of sensors. In this paper we evaluate the performance of a new system for visual control of a selective compliance assembly robot arm (SCARA) robotic arm using a monocular depth map that only requires one monocular camera. This system aims to be an efficient alternative to reduce the number of sensors in the robotic arm area while maintaining the effectiveness of traditional vision algorithms that use stereoscopic architectures of cameras. For this purpose, this system is compared with representative state-of the-art vision algorithms focused on the control of robotic arms. The results are statistically analyzed, indicating that the algorithm proposed in this research has competitive performance compared to state-of-the-art robotic arm visual control algorithms only using a single monocular camera.
Volume: 40
Issue: 2
Page: 801-813
Publish at: 2025-11-01

A systematic evaluation of pre-trained encoder architectures for multimodal brain tumor segmentation using U-Net-based architectures

10.11591/ijeecs.v40.i2.pp850-859
Marwa Abbas , Ashraf A. M. Khalaf , Hussein Mogahed , Aziza I. Hussein , Lamya Gaber , M. Mourad Mabrook
Accurate brain tumor segmentation from medical imaging is critical for early diagnosis and effective treatment planning. Deep learning methods, particularly U-Net-based architectures, have demonstrated strong performance in this domain. However, prior studies have primarily focused on limited encoder backbones, overlooking the potential advantages of alternative pretrained models. This study presents a systematic evaluation of twelve pretrained convolutional neural networks—ResNet34, ResNet50, ResNet101, VGG16, VGG19, DenseNet121, InceptionResNetV2, InceptionV3, MobileNetV2, EfficientNetB1, SE-ResNet34, and SE-ResNet18—used as encoder backbones in the U-Net framework for identification and extraction of tumor-affected brain areas using the BraTS 2019 multimodal MRI dataset. Model performance was assessed through cross-validation, incorporating fault detection to enhance reliability. The MobileNetV2-based U-Net configuration outperformed all other architectures, achieving 99% cross-validation accuracy and 99.3% test accuracy. Additionally, it achieved a Jaccard coefficient of 83.45%, and Dice coefficients of 90.3% (Whole Tumor), 86.07% (Tumor Core), and 81.93% (Enhancing Tumor), with a low-test loss of 0.0282. These results demonstrate that MobileNetV2 is a highly effective encoder backbone for U-Net in extracting tasks for tumor-affected brain regions using multimodal medical imaging data.
Volume: 40
Issue: 2
Page: 850-859
Publish at: 2025-11-01

Generation of distribution routes with shorter distances and fewer vehicles using the simulated annealing algorithm

10.11591/ijeecs.v40.i2.pp707-718
Flor Cardenas-Mariño , Erik Alex Papa Quiroz , Rene Calderon Vilca , Edwar Ilasaca Cahuata , Hesmeralda Rojas Enriquez , Ronald A. Ayquipa Rentería
The vehicle routing problem (VRP) is still a persistent challenge in society, and can be considered a combinatorial optimization problem, where a fleet of delivery vehicles must satisfy the demand of customers sharing the same depot, minimizing the transport distance. The objective of this research is to propose a method to generate distribution routes that minimize both the number of vehicles used and the total distance traveled. To this end, an initial solution is used, on which the Greedy algorithm is applied, followed by the simulated annealing (SA) algorithm, manipulating the exchange techniques, insertion methods, parameter adjustments within the algorithm and applying the penalty as a mechanism to avoid the excessive use of trucks or the assignment of routes that exceed the allowed capacity. The proposal was validated using four datasets, as a result, the general averages of the reduction in distance, changes and penalty cost are shown: The Greedy algorithm reduced the distance by 5.71%, in trucks to 16.57%, in penalty cost to 14.71%; then, applying the SA algorithm, a better efficiency was achieved by reducing the distance by 10.36%, 20.08% in trucks and 18.64% in penalty cost. In this way, the use of vehicles in the distribution routes is optimized, which could contribute to the reduction of vehicular traffic and the reduction of CO2 emissions, thus favoring the environment.
Volume: 40
Issue: 2
Page: 707-718
Publish at: 2025-11-01

Development and evaluation of a generalized ontology framework for software requirement specification

10.11591/ijeecs.v40.i2.pp1050-1064
Sourav Kundu , Soumay Kanti Das , Abu Rafe Md Jamil , Md Kamrul Islam , SK. Shalauddin Kabir , Mostafijur Rahman Akhond
This paper presents an ontology developed to address challenges such as com munication gaps, risks of errors, and inconsistencies during the manual process of creating software requirement specifications (SRS). The proposed ontology offers a systematic and formal depiction of the requirements, enhancing consis tency and communication among stakeholders. The ontology has been devel oped from the software requirements documents to facilitate the development process. This paper discusses the process of creating the ontology and demon strates using Pellet Reasoner for inference and Prot´eg´e for ontology construction to save and reuse information. The ontology seems to be efficient in manag ing complex software projects, enabling accurate requirement retrieval through SPARQL queries. This study emphasizes how incorporating ontologies into re quirement engineering can significantly enhance the quality and reliability of SRS.
Volume: 40
Issue: 2
Page: 1050-1064
Publish at: 2025-11-01

A comparative study of solar photovoltaic array configurations to optimize power harvesting in a real-world system under various partial shading conditions

10.11591/ijeecs.v40.i2.pp558-566
Karthick Balakrishnan , Sudhakaran Mahalingam
Partial shading (PS) significantly reduces power generation and efficiency in solar photovoltaic (PV) systems. This research presents a novel totalcross-tied (TCT) methodology designed to mitigate shading effects by optimizing array layout while preserving electrical connectivity. The TCT method is compared to three established configurations: series-parallel (S-P), bridge-linked (B-L) and honey-comb (H-C). MATLAB simulations on a (9×9) PV array under variousshading conditions demonstrate TCT’s superior performance in achieving the global maximum power point (GMPP). Key findings indicate that TCT surpasses the other configurations, reaching a maximum power output of 16,650W at GMPP, with a mismatch power loss of 2,600W, a power loss of 13.32%, a fill factor (FF) of 38.27, and an execution ratio (ER) of 0.866.
Volume: 40
Issue: 2
Page: 558-566
Publish at: 2025-11-01

Panic detection through facial recognition paradigm using deep learning tools

10.11591/ijeecs.v40.i2.pp1001-1010
Sameerah Faris Khlebus , Mohammed Salih Mahdi , Monji Kherallah , Ali Douik
Recently, panic detection has become essential in security, healthcare, and human-computer interaction. Automatic panic detection (APD) systems are designed to monitor physiological signals and behavioral patterns in real-time to detect stress responses. APD is increasingly adopted across many sectors, including disaster preparedness, COVID-19, and terror attacks. Their integration with various applications reduces human efforts and saves costs. However, most studies rely on existing models with fewer new ones or techniques. This study proposes a vision-based panic detection model using MobileNet, ResNet, and convolutional neural network (CNN). The FER2013 dataset is used for the model training and testing. The results indicate that MobileNet is the most effective model for image-based panic detection across ten folds with an accuracy of 90%, recall of 96.9%, and mean accuracy of 0.032. MobileNet also showed a mean absolute error (MAE) between 0.02 and 0.04. This study has been to confirm MobileNet's suitability for image-based panic detection. The findings contribute to developing more reliable and accurate image-based panic detection systems in real-world applications. It offers valuable insights and lays the groundwork for future deep-leaning-based panic detection studies.
Volume: 40
Issue: 2
Page: 1001-1010
Publish at: 2025-11-01

AlGaN/GaN MSM UV photodetector without and with BGaN back-barrier layer comparison study by SILVACO-TCAD

10.11591/ijeecs.v40.i2.pp590-600
Aicha Benyettou , Abedelkader Hamdoune , Belkacem Benadda , Djamal Lachachi
Using DevEDIT and atlas under SILVCAO-TCAD, we were able to achieve high photodetector metal-semiconductor-metal (MSM) AlGaN/GaN/BGaN performance with high electronic mobility. Our device demonstrated a sensitivity of 286 (I illumination/I dark) at Vanode 20V with an illumination current of 26 mA, a photocurrent of 1.56e-7 A at a wavelength of 0.350 µm, and an appropriate efficiency value of 87% without BGaN, and we also studied the influence of the boron B0.03Ga0.97N back-barrier layer. As a result, we obtain a sensitivity of 293,4 at Vanode 20V with an illumination current of 27 mA, a photocurrent of 1,85e-7 A at a wavelength of 0.350 µm, and an appropriate efficiency value of 90%. Additionally, this type of photodetector has been effectively created to detect UV light in the 100–450 nm range, and it may find value in both medical and military settings. Astronomical, medical diagnostics, environmental sensing, remote sensing, thermal imaging, optical signal detection, night vision cameras, missiles, and target tracking.
Volume: 40
Issue: 2
Page: 590-600
Publish at: 2025-11-01

Interpretable federated deep learning models for predicting gait dynamics in biomechanics

10.11591/ijeecs.v40.i2.pp1087-1099
Shaik Sayeed Ahamed , Akram Pasha , Syed Ziaur Rahman , D. N. Puneeth Kumar
Accurate prediction of human joint angle dynamics and reliable gait classifica tion are essential for applications in rehabilitation, biomechanics, and clinical monitoring. Traditional machine learning (ML) models trained on centralized data raise concerns about privacy, scalability, and transparency. This study proposes a federated deep learning (DL) framework that integrates privacy preserving model training with interpretable predictions. Specifically, a gated recurrent unit- deep neural network (GRU-DNN) hybrid model is developed for regression of joint angles, while a Long short-term memory- convolutional neural network (LSTM-CNN) hybrid model is designed for binary and multi class gait classification. The framework is deployed using the federated av eraging (FedAvg) algorithm across simulated clients, with each client training locally on its data. To enhance interpretability, the local interpretable model agnostic explanations (LIME) algorithm is integrated at the client level to gener ate human-understandable explanations for model predictions. The experimen tal results demonstrate significant improvements, including a reduction in global mean squared error (GMSE) from 56.16 to 3.31 and an increase in R-squared score from 0.80 to 0.99 for regression, along with classification accuracies of 0.97 (binary) and 0.94 (multi-class). This scalable, privacy-preserving frame work bridges the gap between accuracy and transparency, offering impactful applications in biomechanics, healthcare, and personalized medicine.
Volume: 40
Issue: 2
Page: 1087-1099
Publish at: 2025-11-01

Federated learning in edge AI: a systematic review of applications, privacy challenges, and preservation techniques

10.11591/ijeecs.v40.i2.pp926-940
Christina Thankam Sajan , Helanmary M. Sunny , Anju Pratap
Edge artificial intelligence (Edge AI) involves the implementation of AI algorithms and models directly on local edge devices, such as sensors or internet of things (IoT) devices. This allows for immediate processing and analysis of data without the need for continuous dependence on cloud infrastructure. Concerns about privacy have grown importance in recent years for businesses looking to uphold end-user expectations and safeguard business models. Federated learning (FL) has emerged as a novel approach to enhance privacy. To improve generalization qualities, FL trains local models on local data. These models then collaborate to update a global model. Each edge device (like smartphones, IoT sensors, or autonomous vehicles) trains a local model on its own data. This local training helps in capturing data patterns specific to each device or node. Poisoning, backdoors, and generative adversarial network (GAN)-based attacks are currently the main security risk. Nevertheless, the biggest threat to FL’s privacy is from inference-based assaults such as model inversion attacks, differential privacy shortcomings and FL utilizes blockchain and cryptography technologies to improve privacy on edge devices. This paper presents a thorough examination of the current literature on this subject. In more detail, we study the background of FL and its different existing applications, types, privacy threats and its techniques for privacy preservation.
Volume: 40
Issue: 2
Page: 926-940
Publish at: 2025-11-01
1 2 3
...
2097 2098
Show 201 of 2098

Discover Our Library

Embark on a journey through our expansive collection of articles and let curiosity lead your path to innovation.

Explore Now
Library 3D Ilustration