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

30,907 Article Results

MobileNetV2 with transfer learning for brain tumor classification

10.11591/ijeecs.v43.i1.pp281-298
Aziz Srai
This document presents a deep learning-based approach for the automatic classification of brain tumors from magnetic resonance imaging (MRI) images, using the lightweight MobileNetV2 model combined with transfer learning and fine-tuning techniques. The study aims to address the constraints of resource-limited medical environments, where model speed and lightweight design are as important as accuracy. The dataset used is a fusion of three public databases (Figshare, SARTAJ, and BR35H), comprising 4,480 images for training and 1,600 for testing, divided into four classes: glioma, meningioma, pituitary tumor, and healthy brain. The methodology includes several steps: resizing the images to 224×224 pixels, normalizing pixel values between 0 and 1, augmenting the data through random rotations, shifts, and zooms to avoid overfitting, and then extracting features using MobileNetV2 pre-trained on ImageNet. The strategy adopted comprises two phases: first, transfer learning where only the layers added at the top of the model are trained for 10 epochs, then fine-tuning consisting of unfreezing the last 20 layers of the base model and retraining them with a reduced learning rate for 5 epochs. The results obtained show an overall accuracy of 86.56% after fine-tuning, with a macro-mean area under the ROC curve (AUC) of 0.9645, indicating excellent discriminatory power. The confusion matrix reveals that the "no tumor" class achieves perfect performance (400 out of 400), while the "meningioma" class remains the most difficult to classify, often confused with gliomas and pituitary tumors. Compared to more resource-intensive models like VGG16, ResNet50, or EfficientNet, MobileNetV2 offers an optimal balance between performance and lightweight design, with a significantly lower number of parameters, making it particularly well-suited to resource-constrained environments. The authors conclude that this approach provides reliable diagnostic support for radiologists, accelerating tumor detection without replacing medical expertise. Future directions include clinical validation on multi-center data, integration of an automated segmentation step, and exploration of newer architectures such as attention mechanisms.
Volume: 43
Issue: 1
Page: 281-298
Publish at: 2026-07-01

Adaptive vector control of PV-fed induction motor using boost split-source inverter without MPPT

10.11591/ijeecs.v43.i1.pp39-62
Romaissa Hamdi , Yassine Beddiaf , Djamel Sakri , Daoud Rezzak , Hassina Slimani
This paper proposes an adaptive vector control strategy for a photovoltaic (PV)-fed induction motor using a boost split-source inverter (BSSI). Unlike conventional PV conversion systems, the proposed topology combines voltage boosting and DC–AC conversion in a single stage, reducing component count, switching losses, and overall system complexity. In addition, the system operates without a maximum power point tracking (MPPT) algorithm, simplifying the control structure while maintaining stable operation under varying environmental conditions. The proposed control approach integrates sliding mode control (SMC) for robust DC-bus voltage regulation and an adaptive proportional–integral (API) speed controller based on Lie derivative theory for online tuning of controller gains according to the speed tracking error. An active DC-link protection mechanism is also introduced to prevent overvoltage during transient conditions. The main contribution of this paper lies in the combination of the BSSI topology with a hybrid adaptive control framework to improve robustness, dynamic performance, and system reliability under irradiance, temperature, and load variations. Simulation and experimental results demonstrate fast dynamic response, reduced speed and voltage oscillations, accurate speed tracking, and superior performance compared with conventional vector control methods.
Volume: 43
Issue: 1
Page: 39-62
Publish at: 2026-07-01

Pulmonary nodule in CT image quantification by using pulmonary nodules magnitude ratio

10.11591/ijeecs.v43.i1.pp103-113
Asharani Ramadas , Chidananda Murthy Melekote Vinayakamurth
The growth rate of the Pulmonary nodule increases the doubling time of the Pulmonary nodule, which is a significant indicator of malignancy. The mean diameter measurement of the Pulmonary nodule contributes to the assessment of lung cancer based on the doubling time. Small nodule size, partial volume effect, irregular shape, juxtapleural, and juxtavascular nodules are difficult to measure. This work addresses these measurement challenges using an image processing pipeline consisting of image segmentation and quantification. The computed tomography (CT) image preprocessing, segmentation, edge detection, and nodule measurement framework is proposed to extract the nodules from the CT image and then quantify them by measuring their mean diameter. A novel pulmonary nodules magnitude ratio (PNMR) is proposed to establish the sattastical relationship between the nodule and corresponding parenchyma size. The LNMR is evaluated against the synthetic nodules that express the nodule growth rate. This work contributes automatic nodule detection, semiautomatic nodule measurement, and PNMR evaluation for more reliable detection and quantification of Pulmonary nodules.
Volume: 43
Issue: 1
Page: 103-113
Publish at: 2026-07-01

An intelligent speed controller for indirect vector-controlled induction motor with high efficiency taking core loss into account

10.11591/ijeecs.v43.i1.pp28-38
Yassina Mederharhet , Leila Boukarana
To address high-performance drive operations, this work details a genetic algorithm (GA)-tuned proportional integral (PI) control strategy applied to sensorless indirect vector-controlled induction motors (IM), explicitly embedding core loss dynamics within the loop. Although GA-based PI tuning methods have been extensively studied, most existing approaches neglect iron loss dynamics, leading to reduced modeling accuracy and suboptimal energy efficiency. The proposed method simultaneously optimizes PI speed controller gains using GA while integrating core loss resistance into the motor model. This combined optimization enhances both dynamic performance and energy efficiency under varying load and speed conditions. Simulation results demonstrate that the proposed PI-GA controller reduces settling time by 62% compared to a classical PI controller, while overshoot decreases from 18% to 5%. Total harmonic distortion (THD) is limited to 3.4%, and iron losses are reduced by approximately 15%, resulting in an overall efficiency improvement up to 95.2%. Comparative analysis confirms the robustness and superiority of the proposed strategy, highlighting its suitability for high-performance and energy-efficient IM drive applications.
Volume: 43
Issue: 1
Page: 28-38
Publish at: 2026-07-01

Evaluating oversampling methods for imbalanced Arabic dialect identification

10.11591/ijeecs.v43.i1.pp259-270
Maulana Ihsan Ahmad , Aina Musdholifah , Arif Nurwidyantoro
This study investigates whether oversampling is a reliable solution for severe class imbalance in Arabic dialect identification. Using the Shami Corpus as a controlled testbed, we demonstrate that conventional oversampling often fails in high-dimensional sparse text spaces, but density based cluster filtering can effectively resolve this. We conduct a comparative evaluation of SMOTE, clustering-guided variants (ASTRA-SMOTE and SMOTE-RADIANT), and a cost-sensitive ClassWeight approach under an identical 5,644-dimensional feature-engineering pipeline using LightGBM and XGBoost. On the held-out test set, standard SMOTE and class weighting frequently distorted decision boundaries, yielding inconsistent gains across models. In contrast, SMOTE-RADIANT yields a statistically significant macro-F1 improvement for LightGBM (0.8539 vs. 0.8526 on the original data) with a large effect size (r = 0.511), successfully rescuing minority dialects without degrading the majority class. These findings suggest that while oversampling is not universally reliable in sparse text spaces, coupling it with density-based noise neutralization (RADIANT) provides a robust and interpretable alternative to deep learning models. This study provides methodological clarity and reproducible guidance for fair and inclusive Arabic NLP systems.
Volume: 43
Issue: 1
Page: 259-270
Publish at: 2026-07-01

REHA:real-time IoT-based energy efficient home automation system using ESP8266 and PIR motion sensors

10.11591/ijeecs.v43.i1.pp179-191
Md. Kamal Ibne Sufian , Poly Bhoumik , Selina Sharmin , Nazma Tara
Technological developments have improved living standards, leading to greater demand for home automation based on the internet of things (IoT) concept. This study addresses the limitations of existing home automation systems, which are often costly, complex, and lack energy-saving features. A low-cost IoT-based home automation system is developed using the ESP8266 NodeMCU and PIR motion sensors to enable both remote control and automatic operation of house hold appliances. It uses the Blynk platform for cloud-based monitoring with smart motion-based logic to minimize unnecessary power consumption. The evaluation is conducted through various controlled scenarios and an estimate based energy analysis derived from standard appliance ratings. Experimental results showed a reduction in household energy consumption of about 13–15%. The proposed system offers a cost-effective, practical approach to smart home energy management by combining affordability, automation, and measurable efficiency improvements.
Volume: 43
Issue: 1
Page: 179-191
Publish at: 2026-07-01

A mathematical model for IoT malware propagation with adaptive patching strategy based on R₀: a simple optimal control approach

10.11591/ijeecs.v43.i1.pp219-232
Dwi Ely Kurniawan , Sarifuddin Madenda , Eri Prasetyo Wibowo
This paper presents a mathematical framework for modeling and controlling internet of things (IoT) malware propagation via an adaptive patching strategy governed by the real-time basic reproduction number R₀. We introduce the SEIR-P (susceptible–exposed–infected–recovered–patched) model, where the patching control rate u(t) is a sigmoid feedback function of R₀(t). All epidemiological parameters are calibrated from three empirical malware captures of the IoT-23 dataset (Stratosphere laboratory, Czech Technical University) a publicly available labeled collection of real IoT network traffic comprising 23 captures from infected and benign devices: CTU-IoT-1 (Mirai), CTU-IoT-9 (Torii), and CTU-IoT-17 (IRCBot) yielding the first empirically grounded parameter set for SEIR-type IoT epidemic models with confidence intervals. The optimal control problem is formulated via pontryagin’s maximum principle (PMP), and a closed-form R₀(u) expression is derived via the next-generation matrix (NGM), yielding the critical threshold u_crit = 0.142 day⁻¹. Five comparative simulation scenarios over a 365-day horizon show that the proposed R₀-adaptive strategy achieves a 91.3% reduction in peak infection (360 vs. 4,142 devices), eradicates malware by day 179, and attains the highest cost-effectiveness index (CEI = 1.142). Global asymptotic stability of the disease-free equilibrium under u*(t) is proven via Lyapunov’s method and LaSalle’s Invariance Principle. PRCC sensitivity analysis identifies u_max and β as dominant parameters. This closed-loop framework bridges the gap between abstract epidemic theory and deployable IoT security management.
Volume: 43
Issue: 1
Page: 219-232
Publish at: 2026-07-01

Layer-wise adaptive structured pruning via genetic algorithms with taylor-based proxy fitness

10.11591/ijeecs.v43.i1.pp314-324
Anh-Truong Vo , Hoang-Loc Tran , Dinh-Duy Phan , Duc-Lung Vu
Deploying deep convolutional neural networks (CNNs) on edge devices requires balancing model accuracy and computational efficiency. While structured pruning limits inference costs by removing redundant filters, most methods apply a rigid, global criterion, ignoring the distinct representational roles of individual layers. This yields suboptimal results, especially under aggressive compression where over-pruning degrades performance. To address this limitation, we propose an adaptive structured pruning framework based on genetic algorithms (GAs) that jointly optimizes layer-wise pruning ratios and strategies. Each layer independently selects between min-importance and median-rank pruning, enabling the exploration of tailored strategy combinations. A training-free taylor based proxy fitness function ensures efficient candidate evaluation without re peated fine-tuning. After fine-tuning the selected architecture, experiments on VGG16 demonstrate that our method achieves 92.78 ± 0.28% accuracy (over 50 independent runs) with a 70.0 ± 3.2% MACsreductiononCIFAR-10, andmaintains 71.82% accuracy on CIFAR-100. These results demonstrate competitive performance compared to existing pruning methods while achieving substantial computational cost reduction.
Volume: 43
Issue: 1
Page: 314-324
Publish at: 2026-07-01

Hybrid machine learning framework for anomaly detection in industrial IoT environments

10.11591/ijeecs.v43.i1.pp345-354
I Dewa Made Widia , Toni Anwar
The industrial internet of things (IIoT) has become a core component of Industry 4.0, enabling highly connected and data-driven industrial systems while simultaneously increasing exposure to cyber threats. Conventional intrusion detection systems (IDS), especially rule-based and signature-driven approaches, often struggle to cope with the dynamic, high-dimensional, and heterogeneous nature of IIoT traffic. This study proposes a hybrid anomaly detection framework that integrates autoencoder, isolation forest, and long short-term memory (LSTM) models using a weighted decision fusion strategy. Each component contributes complementary capabilities, including nonlinear feature learning, efficient outlier detection, and temporal pattern modeling. The framework is evaluated on the botnet of things (BoT-IoT) dataset and further validated using IoT-23. Experimental results show that the proposed hybrid approach achieves a precision of 0.999, recall of 0.970, and an F1-score of 0.985, while maintaining a false-negative rate below 0.001%. Although its area under the curve (AUC) is slightly lower than that of a standalone light gradient boosting machine (LightGBM) baseline, the hybrid framework consistently reduces missed detections, making it well suited for reliable real-time IIoT security monitoring.
Volume: 43
Issue: 1
Page: 345-354
Publish at: 2026-07-01

A hybrid approach for multi-view MRI Alzehimer’s detection using convolutional neural networks and bio-inspired algorithms

10.11591/ijeecs.v43.i1.pp192-206
Iheb Chemss El Dine Hagani , Nacéra Benamrane , Lakhdar Sais
Alzheimer’s disease (AD) is a neurodegenerative disorder that remains incurable to date. Therefore, the most important step in treatment remains the early detection of the signs indicating its presence. The sooner these signs are discovered, the sooner preventative care can be administered. Convolutional neural networks (CNNs) have demonstrated impressive performance in medical image analysis; however, they often suffer from suboptimal manual tuning of their hyperparameters. Therefore, we opted for a hybrid method combining them with genetic algorithms (GA) and particle swarm optimization (PSO) to automatically optimize architectures and fusion weights for improved AD detection. Using data obtained from ADNI and Kaggle, our approach achieved 87.4% accuracy, surpassing classical CNNs of the same size and depth. These results highlight the potential of evolutionary optimization for developing reliable diagnostic tools.
Volume: 43
Issue: 1
Page: 192-206
Publish at: 2026-07-01

The persistence of simplicity: why naive temporal baselines can outperform machine learning in short-term cancer incidence forecasting

10.11591/ijeecs.v43.i1.pp139-147
Samia Ferhane , Kies Karima
Despite growing enthusiasm for artificial intelligence (AI) in epidemiology, its added value for short-term forecasting remains uncertain. In this method, using CI5plus international cancer registry data (1990–2017) enriched with World Bank urban environment indicators, we benchmark one-year-ahead cancer incidence forecasting under strict temporal validation (train 1990–2012; validation 2013–2015; test 2016–2017). We compare naive temporal base lines (last observation, three-year moving average), classical time-series models (ARIMA, exponential smoothing), machine learning (ridge regression, random forest), and a deep learning model (multilayer perceptron). Model differences are assessed using Diebold–Mariano tests and bootstrap confidence intervals, and robustness is verified through rolling-origin evaluation. Regarding the results, across country–sex–site strata, the last-observation baseline consistently achieved the best test performance, significantly outperforming all competing models (Diebold–Mariano p < 0.05 for all pairwise comparisons on MAE). These results were robust across three rolling-origin windows. Urban environment covariates provided negligible incremental predictive value beyond recent incidence history. In conclusion, for short-horizon cancer incidence forecasting with highly persistent series, strong naive baselines are difficult to beat. Rigorous temporal evaluation, statistical comparison, and appropriate baselines are essential for credible claims of AI benefit in epidemiological forecasting.
Volume: 43
Issue: 1
Page: 139-147
Publish at: 2026-07-01

Analysis of PM2.5 pollutant sources in Jakarta using deep learning models and back trajectory approach

10.11591/ijeecs.v43.i1.pp325-334
Hendro Pratama Saragih , Imas Sukaesih Sitanggang , Hendra Rahmawan
PM2.5 concentrations in Jakarta frequently exceed World Health Organization (WHO) air quality guidelines, indicating the need for an integrated approach for pollution prediction and source assessment. This study develops a spatiotemporal prediction framework using a long short term memory (LSTM) model integrated with the hybrid single particle Lagrangian integrated trajectory (HYSPLIT) model for backward trajectory analysis. Daily PM2.5 data from five monitoring stations were combined with meteorological variables from ERA5, Visualcrossing, and the global data assimilation system, with spatial context evaluated using Sentinel-2 land cover maps. After hyperparameter tuning, the optimized model demonstrated robust predictive capabilities, achieving a peak coefficient of determination (R2) of 75.87% on the test data. The framework exhibited exceptional relative accuracy, particularly at the Jagakarsa and Kebun Jeruk stations, which recorded mean absolute percentage error (MAPE) values of 13.34% and 17.80%, respectively. Backward trajectory analysis during selected pollution episodes indicates two dominant regional transport pathways that may influence PM2.5 levels in Jakarta. These pathways are associated with air mass transport over industrial and built-up areas in eastern and northern regions surrounding Jakarta. Land cover analysis shows limited vegetation along these pathways. Overall, elevated PM2.5 events are associated with combined local emissions, regional transport, and meteorological conditions that limit pollutant dispersion near the surface.
Volume: 43
Issue: 1
Page: 325-334
Publish at: 2026-07-01

Solar photovoltaic power system for Bungin Island

10.11591/ijeecs.v43.i1.pp7-17
Novi Azman , Rudi Naufal Fadhilah , Muhammad Ismail
Bungin Island currently relies on diesel -based electricity power generation, resulting in limited supply reliability. This study evaluates the technical feasibility of an off-grid solar photovoltaic (PV) system integrated with lithium-ion battery energy storage (BESS) using photovoltaic system (PVsyst) simulation. Based on measured load demand of 2,830 kWh/day and NASA solar resource data, the optimized configuration consist of a 0.713 MWp PV array and an 11.54 MWh battery system. Simulation result indicates an annual PV production (E_Array) of 1,258.3 MWh/year. Of this, 1,032.9 MWh/year is delivered to the load, with excess energy of 120.84 MWh/year and a limited unmet load of 13.7 MWh/year (1.33%). The system achieves a solar fraction 98.67%, a performance ratio (PR) of 77.6% and a capacity factor of 17.34%. These results demonstrate that a properly sized PV-battery configuration can reliably replace diesel generation, providing a robust framework for high-renewable electrification in densely populated small islands.
Volume: 43
Issue: 1
Page: 7-17
Publish at: 2026-07-01

Auto-generated unit testing using PCA-aDynaMOSA

10.11591/ijeecs.v43.i1.pp233-249
Made Raja Adi Surya Saputra , Maria Seraphina Astriani
Automated test case generation is essential for improving software quality; however, many-objective search-based testing approaches often experience scalability issues when the number of test objectives increases. This condition leads to slower convergence, higher computational effort, and reduced ability to cover complex program structures. To address this gap, this study proposes an enhanced version of the aDynaMOSA algorithm by incorporating principal component analysis (PCA) to reduce redundant objectives during the search process. The proposed method preserves essential objective information while eliminating dependency noise that typically slows the evolutionary search. Experiments were conducted using the SF110 benchmark dataset through EvoSuite, and the approach was compared with standard many-objective search strategies. The findings demonstrate that PCA-based objective reduction can improve performance, achieving up to 3.87%, 5.50%, and 3.75% for coverage of line, branch, and mutation respectively. These results indicate that dimensionality reduction can significantly enhance scalability and efficiency in automated evolutionary test generation, providing a foundation for future adaptive and hybrid optimization strategies.
Volume: 43
Issue: 1
Page: 233-249
Publish at: 2026-07-01

Development of internet of things-based exoskeleton for monitoring elbow rehabilitation therapy

10.11591/ijres.v15.i2.pp553-562
Geevanthran A/L Vegurgama , Mohd Razali Mohamad Sapiee , Khalil Azha Mohd Annuar
The elbow joint is a complex articulation comprising the humeroulnar and humeroradial joints, facilitating flexion-extension movements essential for daily activities. Leveraging advancements in connected systems and paradigms such as the internet of things (IoT), this study proposes an affordable, effective, and IoT-enabled one-degree-of-freedom (1DOF) elbow exoskeleton for home-based rehabilitation. The exoskeleton is designed to provide a natural range of elbow movements (flexion and extension) while enabling real-time monitoring of rehabilitation progress through mobile applications and web servers. The system collects qualitative data on elbow movements, which are critical in rehabilitation therapy, and enables patients to save their rehabilitation status for future reference. This data can be accessed by doctors remotely, ensuring continuity of care. For patients unable to lift their arm independently, a servomotor provides mechanical assistance, enabling them to achieve desired angles for rehabilitation. The IoT platform generates real-time graphs, offering detailed insights into the recovery process through data analysis. This project is a significant advancement in clinical and healthcare settings, as it reduces dependency on human support or physiotherapists. By integrating IoT technology, the proposed exoskeleton ensures effective, autonomous, and data-driven rehabilitation for elbow joint recovery.
Volume: 15
Issue: 2
Page: 553-562
Publish at: 2026-07-01
1 2 3
...
2060 2061
Show 24 of 2061

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