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

28,451 Article Results

Deep transfer learning for classification of ECG signals and lip images in multimodal biometric authentication systems

10.11591/ijai.v14.i4.pp3160-3171
Latha Krishnamoorthy , Ammasandra Sadashivaiah Raju
Authentication plays an essential role in diverse kinds of application that requires security. Several authentication methods have been developed, but biometric authentication has gained huge attention from the research community and industries due to its reliability and robustness. This study investigates multimodal authentication techniques utilizing electrocardiogram (ECG) signals and face lip images. Leveraging transfer learning from pre-trained ResNet and VGG16 models, ECG signals and photos of the lip area of the face are used to extract characteristics. Subsequently, a convolutional neural network (CNN) classifier is employed for classification based on the extracted features. The dataset used in this study comprises ECG signals and face lip images, representing distinct biometric modalities. Through the integration of transfer learning and CNN classification, improving the reliability and precision of multimodal authentication systems is the primary objective of the study. Verification results show that the suggested method is successful in producing trustworthy authentication using multimodal biometric traits. The experimental analysis shows that the proposed deep transfer learning-based model has reported the average accuracy, F1-score, precision, and recall as 0.962, 0.970, 0.965, and 0.966, respectively.
Volume: 14
Issue: 4
Page: 3160-3171
Publish at: 2025-08-01

Revolutionizing autism diagnosis using hybrid model for autism spectrum disorder phenotyping

10.11591/ijece.v15i4.pp3904-3912
Vijayalaxmi N. Rathod , Rayangouda H. Goudar
The growing prevalence of autism spectrum disorder (ASD) necessitates efficient data-driven screening solutions to complement traditional diagnostic methods, which often suffer from subjectivity and limited scalability. This study introduces a hybrid ensemble model combining logistic regression (LR) and naive Bayes (NB) for ASD classification across four age groups (toddlers, children, adolescents, and adults) using behavioral screening datasets. By integrating statistical learning and probabilistic inference, the proposed model effectively captured behavioral markers, ensuring a higher classification accuracy and improved generalization. The experimental evaluation demonstrated its superior performance, achieving 94.24% accuracy and 99.40% area under the receiver operating characteristic curve (AUROC), surpassing those of individual classifiers and existing approaches. This artificial intelligence (AI)-driven framework offers a scalable, cost-effective, and accessible solution for both clinical and telemedicine-based ASD screening, facilitating early intervention and risk assessment. This study underscores the transformative potential of AI in neurodevelopmental diagnostics, paving the way for more efficient and widely deployable autistic screening technologies.
Volume: 15
Issue: 4
Page: 3904-3912
Publish at: 2025-08-01

Imagery based plant disease detection using conventional neural networks and transfer learning

10.11591/ijai.v14.i4.pp2701-2712
Ali Mhaned , Salma Mouatassim , Mounia El Haji , Jamal Benhra
Ensuring the sustainability of global food production requires efficient plant disease detection, challenge conventional methods struggle to address promptly. This study explores advanced techniques, including convolutional neural networks (CNNs) and transfer learning models (ResNet and VGG), to improve plant disease identification accuracy. Using a plant disease dataset with 65 classes of healthy and diseased leaves, the research evaluates these models' effectiveness in automating disease recognition. Preprocessing techniques, such as size normalization and data augmentation, are employed to enhance model reliability, and the dataset is divided into training, testing, and validation sets. The CNN model achieved accuracies of 95.45 and 94.52% for 128×128 and 256×256 image sizes, respectively. ResNet50 proved the best performer, reaching 98.38 and 98.63% accuracy, while VGG16 achieved 97.99 and 98.34%. These results highlight ResNet50's superior ability to capture intricate features, making it a robust tool for precision agriculture. This research provides practical solutions for early and accurate disease identification, helping to improve crop management and food security.
Volume: 14
Issue: 4
Page: 2701-2712
Publish at: 2025-08-01

The growth and trends information technology endangered language revitalization research: Insight from a bibliometric study

10.11591/ijece.v15i4.pp3888-3903
Leonardi Paris Hasugian , Syifaul Fuada , Triana Mugia Rahayu , Apridio Edward Katili , Feby Artwodini Muqtadiroh , Nur Aini Rakhmawati
Since United Nations Educational, Scientific and Cultural Organization (UNESCO) declared endangered languages, researchers have revitalized endangered languages in many fields. This study discusses a bibliometric analysis conducted to investigate research on the topic of revitalization of endangered languages in information technology. The study's aim is to assess research topics by identifying authors, institutions, and countries that influence research collaboration. The Scopus dataset (from 2002-2024) was obtained from journal articles (n=62) and conference papers (n=76) and visualized using VOSviewer 1.6.20. The analysis outcomes reveal a fluctuating trend with an increasing pattern. The United States, Canada, and China were identified as the top three countries in terms of publications. Meanwhile, the University of Alberta, Université du Québec à Montréal, University of Auckland, and University of Hawaiʻi at Mānoa are the most prolific institutions on this topic, with two authors from the Université du Québec à Montréal, Sadat and Le, being the most productive. The dominant research is related to computational linguistics. Meanwhile, topics such as phonetic posteriograms, integrated frameworks, and artificial intelligence are some of the potential research areas that can be explored in the future. Its implications for exposing the extent to which the development of endangered language revitalization can be accommodated in the field of information technology.
Volume: 15
Issue: 4
Page: 3888-3903
Publish at: 2025-08-01

Optimizing convolutional neural network hyperparameters to enhance liver segmentation accuracy in medical imaging

10.11591/ijece.v15i4.pp3876-3887
Iwan Purnama , Agus Perdana Windarto , Solikhun Solikhun
Liver segmentation in medical imaging is a crucial step in various clinical applications, such as disease diagnosis, surgical planning, and evaluation of response to therapy, which require a high degree of precision for accurate results. This research focuses on increasing the accuracy of liver segmentation by optimizing hyperparameters in the convolutional neural network (CNN) model using the developed ResNet architecture. The uniqueness of this research lies in the application of hyperparameter optimization methods such as random search and Bayesian optimization, which allow broader and more efficient exploration than conventional approaches. The results show that the DeepLabV3Plus model (the proposed model) significantly outperforms the standard ResNet in the image segmentation task. DeepLabV3Plus shows excellent performance with an MIoU score of 0.965, a PA Score of 0.929, and a meager loss value of 0.011. These results show that DeepLabV3Plus is able to recognize and predict segmentation areas very accurately and consistently and minimize prediction errors effectively. In conclusion, the results of this study show a significant improvement in segmentation accuracy, with the optimized model providing better performance in the evaluation.
Volume: 15
Issue: 4
Page: 3876-3887
Publish at: 2025-08-01

Development and evaluation of a smart home energy management system using internet of things and real-time monitoring

10.11591/ijece.v15i4.pp3977-3985
Mohamed Imran Mohamed Ariff , Nur Anim Abdul Halim , Mohammad Nasir Abdullah , Samsiah Ahmad , Masurah Mohamad , Anis Zafirah Azmi
This project presents the design and implementation of a smart home energy management system using internet of things (IoT) technology to optimize household energy consumption. The system integrates various sensors, including passive infrared (PIR), light dependent resistor (LDR), and DHT11, to collect real-time environmental data, which is processed by a NodeMCU microcontroller. The microcontroller controls home appliances using relays, while the Blynk mobile app and Streamlit web platform provide users with remote monitoring and control capabilities. Despite successfully optimizing energy usage, the system faces limitations such as high sensor sensitivity and potential hazards during high-load power demonstrations. To address these issues, future work proposes integrating additional sensors for improved accuracy and incorporating renewable energy sources for increased sustainability. This project aims to enhance energy efficiency, provide users with greater control over their energy consumption, and contribute to smart home automation by utilizing real-time data, IoT integration, and user-friendly interfaces.
Volume: 15
Issue: 4
Page: 3977-3985
Publish at: 2025-08-01

Blockchain as a digital governance tool: A systematic review

10.11591/ijece.v15i4.pp3986-3995
Cesar Patricio-Peralta , Jimmy Ramirez Villacorta , Milton Amache Sánchez , Jacker Paredes Meneses , Jesús Zamora Mondragon , Luis Segura Terrones , Paul Torres Santos , César Veliz Manrique , Walter Patricio Peralta
This systematic review explores the implementation of blockchain technology as a digital governance tool, focusing specifically on the Peruvian context. In the digital transformation era, blockchain has established itself as an innovative solution to manage and authenticate information. This research focuses on optimizing administrative and governmental processes in Peru, a country where document verification is crucial in legal, financial, educational, and medical procedures. The methodology used follows the problem/population, intervention, comparison, outcome, context (PICOC) model. 56 high-impact articles were selected in Scopus, prioritizing those in the areas of engineering, computer science, and business, and published between 2022 and 2025. The objective was to define the scope and structure of the research questions. These questions address the implementation of blockchain and its applications in digital governance to ensure security and reliability in administrative procedures. Through a comprehensive literature review, we seek to provide a comprehensive view of how blockchain could transform the interaction between citizens and the Peruvian government by automating document verification. In addition, successful cases from other countries and similar sectors will be analyzed, evaluating their feasibility and applicability in the Peruvian context. This approach will allow us to identify both the potential benefits and the challenges and implications associated with the integration of blockchain into government processes in Perú.
Volume: 15
Issue: 4
Page: 3986-3995
Publish at: 2025-08-01

Navigating cyber investigations: strategies and tools for forensic data acquisition

10.11591/ijece.v15i4.pp4022-4030
Srinivas Kanakala , Vempaty Prashanthi , K. V. Sharada
The rapid proliferation of cybercrimes has underscored the critical importance of robust data acquisition methodologies in the field of digital forensics. This research publication explores various aspects of forensic data acquisition, focusing on techniques, tools, and best practices employed by forensic investigators to collect and preserve digital evidence effectively. Beginning with an overview of the escalating cyber threat landscape and the consequential need for forensic investigations, the publication delves into the fundamental concepts of data acquisition, emphasizing the significance of ensuring data integrity and admissibility in legal proceedings. It examines the process of acquiring both volatile and non-volatile data from diverse sources, including hard drives, RAM, and other digital storage media. Furthermore, evaluates a range of forensic imaging and validation methods, encompassing tools such as Belkasoft live RAM capturer, AccessData FTK Imager, and ProDiscover, alongside validation techniques using PowerShell utility and commercial forensic software. Through comprehensive analysis and discussion, this study serves as a valuable resource for forensic practitioners, researchers, and legal professionals seeking to enhance their understanding of forensic data acquisition methodologies in the ever-evolving landscape of cybercrime investigation.
Volume: 15
Issue: 4
Page: 4022-4030
Publish at: 2025-08-01

Optimization model of vehicle routing problem with heterogenous time windows

10.11591/ijece.v15i4.pp4043-4057
Herman Mawengkang , Muhammad Romi Syahputra , Sutarman Sutarman , Gerhard Wilhelm Weber
This study proposes a novel optimization framework for the vehicle routing problem with heterogeneous time windows, a critical aspect in logistics and supply chain operations. Unlike conventional vehicle routing problem (VRP) models that assume uniform service schedules and fleet capacities, our approach acknowledges the diverse time constraints and vehicle specifications often encountered in real-world scenarios. By formulating the problem as a mixed integer linear programming model, we incorporate constraints related to time windows, vehicle load capacities, and travel distances. To tackle the NP-hard complexity, we employ a hybrid strategy combining metaheuristic algorithms with exact methods, thus ensuring both solution quality and computational efficiency. Extensive computational experiments, conducted on benchmark datasets and real-world logistics data, confirm the superiority of our model in terms of solution quality, runtime, and adaptability. These findings underscore the model’s practicality for industries facing dynamic routing requirements and tight service windows. Furthermore, the proposed framework equips decision-makers with a robust tool for optimizing route planning, ultimately enhancing service quality, reducing operational costs, and promoting more reliable delivery outcomes.
Volume: 15
Issue: 4
Page: 4043-4057
Publish at: 2025-08-01

Prediction of side effects of drug resistant tuberculosis drugs using multi-label random forest

10.11591/ijai.v14.i4.pp2899-2908
Siti Syahidatul Helma , Wisnu Ananta Kusuma , Mushthofa Mushthofa , Diah Handayani
Drug-resistant tuberculosis (DR-TB) has become a concern because anti-tuberculosis drugs (ATD) used to treat it can cause side effects in patients. This study aimed to predict the potential side effects of ATD using a multi-label classification approach with a random forest (RF) algorithm. This study used 660 medical record data, including the 14 ATD treatments prescribed to the patients and the six side effects experienced by patients. The model was trained using the best parameters based on the hyperparameter tuning process. The results show that the RF multi-label algorithm can be an alternative for building ATD side effect prediction models because it produces the most optimal performance value compared to the decision tree (DT) and extreme gradient boosting (XGBoost). The area under the curve (AUC) score of all RF multi-label models is above 0.8, which means that all RF multi-label models are considered acceptable and applicable for ATD side effect prediction. In addition, eight features influenced the models based on the average feature importance score of the RF models. This study is expected to help predict the side effects of ATD used to treat DR-TB based on ATD treatment and determine the most promising tree-based machine learning algorithm for predicting ATD side effects.
Volume: 14
Issue: 4
Page: 2899-2908
Publish at: 2025-08-01

Using the ResNet-50 pre-trained model to improve the classification output of a non-image kidney stone dataset

10.11591/ijai.v14.i4.pp3182-3191
Kazeem Oyebode , Anne Ngozi Odoh
Kidney stone detection based on urine samples seems to be a cost-effective way of detecting the formation of stones. Urine features are usually collected from patients to determine if there is a likelihood of kidney stone formation. There are existing machine learning models that can be used to classify if a stone exists in the kidney, such as the support vector machine (SVM) and deep learning (DL) models. We propose a DL network that works with a pre-trained (ResNet-50) model, making non-image urine features work with an image-based pre-trained model (ResNet-50). Six urine features collected from patients are projected onto 172,800 neurons. This output is then reshaped into a 240 by 240 by 3 tensors. The reshaped output serves as the input to the ResNet-50. The output of this is then sent into a binary classifier to determine if a kidney stone exists or not. The proposed model is benchmarked against the SVM, XGBoost, and two variants of DL networks, and it shows improved performance using the AUC-ROC, Accuracy and F1-score metrics. We demonstrate that combining non-image urine features with an image-based pre-trained model improves classification outcomes, highlighting the potential of integrating heterogeneous data sources for enhanced predictive accuracy.
Volume: 14
Issue: 4
Page: 3182-3191
Publish at: 2025-08-01

Revolutionizing internet of things intrusion detection using machine learning with unidirectional, bidirectional, and packet features

10.11591/ijai.v14.i4.pp3047-3062
Zulhipni Reno Saputra Elsi , Deris Stiawan , Bhakti Yudho Suprapto , M. Agus Syamsul Arifin , Mohd. Yazid Idris , Rahmat Budiarto
Detection of attacks on internet of things (IoT) networks is an important challenge that requires effective and efficient solutions. This study proposes the use of various machine learning (ML) techniques in classifying attacks using unidirectional, bidirectional, and packet features. The proposed methods that implement decision tree (DT), random forest (RF), extreme gradient boosting classifier (XGBC), AdaBoost (AB) and linear discriminant analysis (LDA) work perfectly with all kinds of datasets and includes. It also works very well with data type-based feature selection (DTBFS) and correlation-based feature selection (CBFS). The experiment results show a significant improvement compared to previous studies and reveals that unidirectional and bidirectional features provide higher accuracy compared to packet features. Furthermore, ML models, particularly DT, and RF, have faster computing times compared to more complex deep learning models. This analysis also shows potential overfitting in some models, which requires further validation with different datasets. Based on these findings, we recommend the use of RF and DT for scenarios with unidirectional and bidirectional features, while AB and LDA for packet features. The study concludes that using the right ML techniques along with features that work in both directions can make an intrusion detection system for IoT networks becomes very accurate.
Volume: 14
Issue: 4
Page: 3047-3062
Publish at: 2025-08-01

Integrating time-frequency features with deep learning for lung sound classification

10.11591/ijece.v15i4.pp3737-3747
Su Yuan Chang , Marni Azira Markom , Zhi Sheng Choong , Arni Munira Markom , Latifah Munirah Kamaruddin , Erdy Sulino Mohd Muslim Tan
Deep learning has transformed medical diagnostics, especially in analyzing lung sounds to assess respiratory conditions. Traditional methods like CT scans and X-rays are impractical in resource-limited settings due to radiation exposure and time consumption, while conventional stethoscopes often lead to misdiagnosis due to subjective interpretation and environmental noise. This study evaluates deep learning models for lung sound classification using the International Conference on Biomedical Health Informatics 2017 dataset, comprising 920 annotated samples from 126 subjects. Pre-processing includes down sampling, segmentation, normalization, and audio clipping, with feature extraction techniques like spectrogram and Mel-frequency cepstral coefficients (MFCC). The adopted automatic lung sound diagnosis network (ASLD-Net) model with triple feature input (time domain, spectrogram, and MFCC) achieved the highest accuracy at 97.25%, followed by the dual feature model (spectrogram and MFCC) at 95.65%. Single-input models with spectrogram and MFCC performed well, while the time domain input alone had the lowest accuracy.
Volume: 15
Issue: 4
Page: 3737-3747
Publish at: 2025-08-01

Evaluation of the dynamic performance and practical limitations of a two-wheeled self-balancing robot

10.11591/ijece.v15i4.pp3613-3620
Rupasinghe Arachchige Don Dhanushka Dharmasiri , Malagalage Kithsiri Jayananda
Two-wheeled self-balancing robots (TWSBR) are statically unstable. However, using closed-loop controllers can stabilize. In this work, the proportional-integral-derivative (PID) controller was designed to maintain the TWSBR stability by adding two zeros and a pole at the origin to the loop gain and by determining the parameter K via root-locus analysis. Then using the K value Kp, Ki, and Kd parameters were calculated. By applying an impulse response to the system, it was found that the system is able to reach a dynamic balance in less than 1.2 seconds with minimum steady-state error. The dynamic performance and limitations of the developed system were investigated. The highest disturbance angle that can be applied to the system while keeping the motor input voltage below 12 V, in order to create counterbalancing torque and achieve dynamic balance, is determined to be θ = 0.0524 rad. Additionally, it was found that the TWSBR system managed to retain stability in a significantly large range of sudden payload changes with the same PID controller.
Volume: 15
Issue: 4
Page: 3613-3620
Publish at: 2025-08-01

Optimized reactive power management system for smart grid architecture

10.11591/ijece.v15i4.pp3707-3716
Manju Jayakumar Raghvin , Manjula R. Bharamagoudra , Ritesh Dash
The Indian power grid is an extensive and mature power system that transfers large amounts of electricity between two regions linked by a power corridor. The increased reliance on decentralized renewable energy sources (RESs), such as solar power, has led to power system instability and voltage variations. Power quality and dependability in a smart grid (SG) setting can be enhanced by the careful tracking and administration of solar energy generated by panels. This study proposes a number of reactive power regulation algorithms that take smart grids into account. When developing a kernel, debugging is a must in optimal reactive power management. In this research, a debugging primitive called physical memory protection (PMP), a security feature, is considered. Debugging in the kernel domain requires specialized tools, in contrast to the user space where we have kernel assistance. This research proposes an optimal reactive power management in smart grid using kernel debugging model (ORPM-SG-KDM) for managing the reactive power efficiently. This research achieved 98.5% accuracy in kernel debugging and 99.2% accuracy in optimal reactive power management. Kernel debugging accuracy is increased by 1.8% and 3% of reactive power management accuracy is increased.
Volume: 15
Issue: 4
Page: 3707-3716
Publish at: 2025-08-01
1 2 3
...
1896 1897
Show 64 of 1897

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