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

30,938 Article Results

Advances in dermatological imaging: enhancing skin melanoma classification for improved patient outcomes

10.11591/csit.v7i1.p111-120
Debadutta Sahoo , Soumya Mishra
The study presents an enhanced AlexNet-based deep learning system for binary classification of melanoma skin cancer as either benign or malignant using two paired dermatoscopic and clinical image datasets. The study evaluates the resilience of the models across different image sets with common preprocessing and specific data augmentation, using a melanoma dataset containing 10,000 images and a benign versus malignant dataset with 3,600 images. The AlexNet refinement exceeded several standard machine learning (ML) classifiers and other deep architectures on the two datasets with practical training times, gaining 97.12% and 96.21% in balanced accuracy. The training proceeded with SGD as optimiser and cross-entropy as loss on 256×256 images. Benchmarking against support vector machine (SVM), k-nearest neighbour (KNN), and other convolutional neural networks (CNNs) designs shows that the selected architecture and hyperparameters achieved the highest performance on cost-effective computation for the routine melanoma triage. The report highlights the need for external validation, incorporation into dermatological workflows, and explainability to improve trust, diminish dataset bias, and support the safe clinical deployment in practice.
Volume: 7
Issue: 1
Page: 111-120
Publish at: 2026-03-01

Raindrop and bit drop effects on millimeter wave network performance: a critical review

10.11591/csit.v7i1.p83-92
Victor Dela Gordon , Amevi Acakpovi , George Kwamena Aggrey , Michael Gameli Dziwornu
This PRISMA guided review examines how rain precipitation degrades 5G millimeter wave (mmWave) network performance, with emphasis on rain induced bit drop and its impact on end-to-end quality of service (QoS). From an initial corpus of 13,317 publications screened across IEEE Xplore, ACM Digital Library, ScienceDirect, Google Scholar, and ELICIT, 18 peer reviewed studies published between 2018 and 2024 met the inclusion criteria. Findings show that rainfall significantly weakens mmWave signals, with specific attenuation ranging from approximately 4 to 45 dB/km at 100 mm/h, particularly in tropical regions. When QoS outcomes are reported, these losses manifest as increased bit error rates, rain driven bit drop along the link, higher packet loss and delay, and reduced throughput. Key deficiencies identified include limited empirical validation of attenuation models against packet level QoS, lack of standardized propagation datasets for short range links, and weak treatment of bit level impairments within QoS analysis. To address these gaps, the review recommends enhancing ITU R P.530 and Mie scattering models with region specific measurements, implementing rain aware adaptive protocols, and adopting standardized benchmarking frameworks that link rain attenuation, bit drop, and QoS. This synthesis offers guidance for building climate aware mmWave systems and positions bit drop as a practical metric for precipitation resilience assessment.
Volume: 7
Issue: 1
Page: 83-92
Publish at: 2026-03-01

Implementation of face recognition using Python

10.11591/csit.v7i1.p1-9
Febrian Wahyu Christanto , Husnul Arifin , Christine Dewi , Teguh Prasandy
Artificial intelligence (AI)-based technology systems are developing rapidly. Along with technological development the number of criminal cases caused by facial forgery is also growing. Cases of theft and housebreaking with fake photos are a common problem in Semarang. In 2022–2023 the number of cases of theft and housebreaking reached 372,965 with a crime risk level of 137/100,000 people. To overcome this problem the facial recognition system used in the door security system uses digital image processing. This method works by imitating how nerve cells communicate with interconnected neurons, or more precisely, how artificial neural networks function in humans. As training data, image capture and facial recognition are carried out using a webcam and the Python programming language with the TensorFlow library. The image processing algorithm uses 400 facial images with an accuracy rate of 95%. However further development is needed to improve the efficiency and accuracy of the system to produce better results.
Volume: 7
Issue: 1
Page: 1-9
Publish at: 2026-03-01

Advances in Parkinson’s disease diagnosis and treatment using artificial intelligence: a review

10.11591/csit.v7i1.p121-130
Mehr Ali Qasimi , Züleyha Yılmaz Acar
Parkinson’s disease (PD) diagnosis and monitoring have significantly improved because to current advancements in artificial intelligence (AI), particularly in the areas of deep learning (DL) and machine learning (ML). Early-stage insensitivity of traditional diagnostic techniques necessitates the use of clever, data-driven alternatives. AI-powered noninvasive diagnostic methods like speech recognition, handwriting analysis, and neuroimaging categorization are the main topic of this technical review. We provide a summary of comparative performance measures from recent models, highlighting their practical usefulness, data modality, and accuracy. Also covered are important issues like data variability, real-world implementation, and model interpretability. Unlike prior surveys that primarily report accuracy metrics, this review explicitly focuses on identifying the gap between experimental AI performance and real-world clinical deployment, emphasizing interpretability, validation, and scalability challenges in PD diagnosis. The purpose of this letter is to provide guidance for researchers creating deployable and clinically valid AI systems for PD detection.
Volume: 7
Issue: 1
Page: 121-130
Publish at: 2026-03-01

An uneven cluster-based routing protocol for WSNs using a hybrid MCDM and max-min ant colony optimization

10.11591/csit.v7i1.p74-82
Man Gun Ri , Pyong Gwang Kim , JinSim Kim
In energy-constrained wireless sensor networks (WSNs) composed of sensor nodes (SNs) characterized by multi-criteria contradictory with each other, it is still one of the challenges to be solved to figure out how to combine multi-criteria with each other and how to use an intelligent optimization (IO) algorithm for developing an optimal cluster-based routing protocol. In this article, we overture a new routing protocol based on uneven cluster using the hybrid FCNP-VWA-TOPSIS (FVT) and an improved max-min ant colony optimization (ACO). This scheme uses the hybrid FVT to perform the clustering, and uses an improved max-min ACO to configure a routing tree for the relay transmission of sensed data. The extensive simulation experiments have been carried out to show that the proposed scheme greatly prolongs the network lifetime (NL) by achieving an energy consumption balance superior to the previous schemes.
Volume: 7
Issue: 1
Page: 74-82
Publish at: 2026-03-01

Optimizing interconnection call routing: a machine learning approach for cost and quality efficiency

10.11591/csit.v7i1.p56-65
Ivy Anesu Mudari , Mainford Mutandavari , Kenneth Chiworera
This study presents the design and development of an automated least cost routing (LCR) model for telecommunications interconnection calls using machine learning. Leveraging a random forest regressor, the model predicts the most cost-effective call routing path based on pricing and network latency. Trained on real-world call detail records (CDRs) from TelOne Zimbabwe, the model achieved a high R² score of 0.851, with a mean absolute error (MAE) of $0.0482 per minute. Evaluation results demonstrate an average cost reduction of 46.75% compared to traditional routing methods, with prediction times under 0.1 seconds and latency remaining within acceptable thresholds. This work provides a practical, scalable, and efficient solution for telecom. operators seeking to reduce interconnection costs and maintain service quality through intelligent routing automation. The model architecture and performance to make it viable for integration into real-time telecom infrastructure.
Volume: 7
Issue: 1
Page: 56-65
Publish at: 2026-03-01

Development and performance evaluation of a CNN model for seagrass species classification in Bintan, Indonesia

10.11591/csit.v7i1.p20-29
Nurul Hayaty , Hollanda Arief Kusuma
This study presents the development and evaluation of a convolutional neural network (CNN) model for automated seagrass species classification in Bintan, Indonesia. The objective of this research is to examine how different train-validation data split ratios affect model accuracy and generalization performance. The CNN was trained under four configurations (60:40, 70:30, 80:20, and 90:10) to analyze the influence of training data volume on learning convergence and predictive capability. The results indicate that all configurations achieved high validation accuracy, with the best performance reaching 98.53% when using the 90:10 split. Evaluation on unseen data demonstrated that the 60:40 configuration provided the most consistent and reliable generalization. Performance variations were also affected by the morphological similarity between the classified species, which increases the challenge in correctly distinguishing certain classes. Overall, the findings confirm the effectiveness of CNN-based classification for supporting marine biodiversity monitoring and underline the importance of dataset composition in achieving optimal performance. Future improvements will focus on expanding data variability to enhance robustness in real-world scenarios.
Volume: 7
Issue: 1
Page: 20-29
Publish at: 2026-03-01

Cloud-based predictive analytics for pension fund performance optimization

10.11591/csit.v7i1.p46-55
Beauty Garaba , Mainford Mutandavari , Jerita Chibhabha
This study introduces a novel, cloud-based predictive analytics framework tailored for pension fund performance management in Zimbabwe. Addressing limitations in traditional actuarial models, the proposed system leverages real-time data pipelines and explainable artificial intelligence (XAI) techniques to enhance forecasting accuracy and transparency. Using regression, classification, and deep learning models, it forecasts member contributions, identifies risks of contribution drops, and predicts member churn. The system’s cloud deployment ensures scalability and interactive integration with tools like Power BI for decision support. This solution significantly advances sustainable pension fund management for emerging economies.
Volume: 7
Issue: 1
Page: 46-55
Publish at: 2026-03-01

Bridging archaeological visibility analysis and real-time 3D visualization

10.11591/csit.v7i1.p93-101
George Malaperdas , Georgia Delli
This paper investigates the integration of geographic information systems (GIS)-based visibility analysis—commonly known as viewshed analysis—with real-time 3D rendering in unreal engine, specifically within the context of archaeological and cultural heritage applications. Visibility maps are an essential tool in archaeological research, helping scholars understand the spatial relationships, sightlines, and symbolic visibility between structures, monuments, and landscapes. However, traditional GIS viewshed analysis is often static and limited to 2D environments. This project proposes a method to bring visibility analysis into immersive 3D environments by visualizing GIS-generated data within unreal engine. The methodology involves generating a viewshed from a given digital elevation model (DEM) using established GIS software. The resulting raster is then exported and processed into a texture or material mask compatible with unreal engine. Once imported, the data is mapped onto a 3D landscape model, allowing users to explore visibility dynamically, including first-person or VR-based navigation. This interdisciplinary approach contributes to the field of digital archaeology by enhancing spatial interpretation and audience engagement through immersive geovisualization. It also outlines a flexible pipeline for integrating geospatial datasets into 3D environments, potentially applicable to site management, public education, and digital preservation efforts.
Volume: 7
Issue: 1
Page: 93-101
Publish at: 2026-03-01

Car selection in games using multi-objective optimization by ratio analysis based on player achievement

10.11591/csit.v7i1.p30-45
Caesar Nafiansyah Putra , Fresy Nugroho , Mochamad Imamudin , Dwi Pebrianti , Jehad Abdelhamid Hammad , Tri Mukti Lestari , Dian Maharani , Alfina Nurrahman
The selection menu in some racing games usually uses a random system for vehicle selection. However, this random feature generally randomizes the selection of the index without considering factors that support the player's abilities. Therefore, this study aims to develop a racing game that can suggest vehicles that have been adjusted to the player's performance. Vehicle recommendations are made using the multi-objective optimization on the basis of ratio analysis (MOORA) method as its method. The MOORA calculation ranks vehicles based on criteria such as mileage, fuel efficiency, speed, agility, and others collected in previous games. The results of this study show the effectiveness of using the MOORA method in recommending vehicles that match the player's skills, thereby improving the overall player experience. In addition, the usability test produced a system usability scale (SUS) score of 82.4, so it is included in the very good category.
Volume: 7
Issue: 1
Page: 30-45
Publish at: 2026-03-01

Deep learning for sentiment analysis and topic extraction in health insurance

10.11591/csit.v7i1.p66-73
Muzondiwa Karomo , Mainford Mutandavari , Wilton Muzava
Social media has transformed into a vital channel for real-time, unsolicited feedback in healthcare, yet health insurance providers often lack the tools to mine insights from such data. This study proposes a cloud-based system leveraging deep learning for sentiment analysis and topic modeling tailored to the Commercial and Industrial Medical Aid Society (CIMAS) health insurance in Zimbabwe. Using bidirectional encoder representations from transformers (BERT), a convolutional neural network (CNN), a random forest (RF), and autoencoders, the system processes multilingual data from platforms like Twitter and Facebook, identifying customer concerns in real time. Over 15,000 posts were analyzed, with CNN achieving 91.4% accuracy in sentiment classification and BERTopic extracting coherent themes. The system detected issues such as claim delays, app navigation problems, and unreported anomalies. Findings demonstrate that AI can improve service delivery, customer satisfaction, and responsiveness in African insurance contexts.
Volume: 7
Issue: 1
Page: 66-73
Publish at: 2026-03-01

Enhanced smart farming security with class-aware intrusion detection in fog environment

10.11591/ijict.v15i1.pp257-266
Selvaraj Palanisamy , Radhakrishnan Rajamani , Prabakaran Pramasivam , Mani Sumithra , Prabu Kaliyaperumal , Rajakumar Perumal
The adoption of the internet of things (IoT) in smart farming has enabled real-time data collection and analysis, leading to significant improvements in productivity and quality. However, incorporating diverse sensors across large-scale IoT systems creates notable security challenges, particularly in dynamic environments like Fog-to-Things architectures. Threat actors may exploit these weaknesses to disrupt communication systems and undermine their integrity. Tackling these issues necessitates an intrusion detection system (IDS) that achieves a balance between accuracy, resource optimization, compatibility, and affordability. This study introduces an innovative deep learning-driven IDS tailored for fog-assisted smart farming environments. The proposed system utilizes a class-aware autoencoder for detecting anomalies and performing initial binary classification, with a SoftMax layer subsequently employed for multi-class attack categorization. The model effectively identifies various threats, such as distributed denial of service (DDoS), ransomware, and password attacks, while enhancing security performance in environments with limited resources. By utilizing the Fog-to-Things architecture, the proposed IDS guarantees reliable and low-latency performance under extreme environmental conditions. Experimental results on the TON_IoT dataset reveal excellent performance, surpassing 98% accuracy in both binary and multi-class classification tasks. The proposed model outperforms conventional models (convolutional neural network (CNN), recurrent neural network (RNN), deep neural network (DNN), and gated recurrent unit (GRU)), highlighting its superior accuracy and effectiveness in securing smart farming networks.
Volume: 15
Issue: 1
Page: 257-266
Publish at: 2026-03-01

Review on patch antenna for 5G Networks at Ka-Band

10.11591/csit.v7i1.p102-110
Md. Nurullah Al Nasib , Md. Sohel Rana
Microstrip antennas for Ka-band wireless applications will be thoroughly examined in this research. To utilize 5G wireless applications, a new research topic that has been established is the creation of microstrip patch antennas. Patch antennae are made of different shapes, such as rectangles, circular shapes, triangles, donuts, rings, etc. Many substrate materials are used in patch antenna designs. This article examines the geometric configurations of antennas, the many methods of analysis for attributes of antennas, the dimensions of antennas, the issues that antennas face, and the potential solutions to those challenges. Wireless communication technologies, such as television broadcasts, microwave ovens, mobile phones, wireless local area networks (LANs), Bluetooth, global positioning systems (GPS), and two-way radios, all use it. This article examines the geometric structures of antennas, including several characteristics and materials by which they are constructed, as well as the numerous shapes they can produce. This paper will also examine return loss (S11), bandwidth, voltage standing wave ratio (VSWR), gain, directivity, efficiency, and Bandwidth discussed in the prior studies. In the future, a novel patch antenna can be designed for 5G wireless applications.
Volume: 7
Issue: 1
Page: 102-110
Publish at: 2026-03-01

Optimizing solar energy forecasting and site adjustment with machine learning techniques

10.11591/ijict.v15i1.pp384-392
Debani Prasad Mishra , Jayanta Kumar Sahu , Soubhagya Ranjan Nayak , Anurag Panda , Priyanshu Paramjit Dash , Surender Reddy Salkuti
Estimation of solar radiation is a key task in optimizing the operation of power systems incorporating high levels of photovoltaic (PV) generation. This paper discusses the application of machine learning techniques, namely extreme gradient boosting (XGBT) and random forest (RF), to improve accuracy in the forecasting of solar radiation while adapting for different sites. Utilizing datasets such as meteorological and solar radiation data, the suggested models demonstrate the enhancement of forecasting accuracy by 39% from traditionally applied statistical practices. Along with this, this study also encompasses how endogenous and exogenous factors could be involved in better predictions of solar energy availability. From our findings, XGBT, as well as other machine learning techniques, do enjoy superior performance levels when it comes to the forecasting of solar radiation, which in turn promotes efficient management and potential adaptation of solar energy systems. This study demonstrates how this last generation of algorithms could be applied to noticeably improve the efficiency of solar power forecasting and thereby contribute to more sustainable and reliable energy systems as a byproduct of that.
Volume: 15
Issue: 1
Page: 384-392
Publish at: 2026-03-01

DeepRetina: a multimodal framework for early diabetic retinopathy detection and progression prediction

10.11591/ijict.v15i1.pp152-160
Sunder Ramasamy , Brindha Mohanraj , Sridhar Pushpanathan , Thenmozhi Elumalai , Prabu Kaliyaperumal , Rajakumar Perumal
Diabetic retinopathy (DR) remains one of the top causes of vision loss globally, and early detection and accurate progression prediction are critical in its management. This paper introduces DeepRetina, a deep learning framework that integrates state-of-the-art multimodal retinal imaging techniques with patient-specific clinical data for the improved diagnosis and prognosis of DR. DeepRetina harnesses cutting-edge convolutional neural networks (CNNs) and attention mechanisms to jointly analyze optical coherence tomography (OCT) scans and fundus photographs. The architecture further includes a temporal module that investigates the longitudinal changes in the retina. DeepRetina fuses these heterogeneous data sources with patient clinical information in pursuit of early detection of DR and provides personalized predictions for the progression of the disease. We use a specially designed CNN architecture to process high-resolution retinal images, coupled with a self-attention mechanism that focuses on the most relevant features. This recurrent neural network (RNN) module empowers it to integrate time-series data that captures the evolution of retinal abnormalities. Another neural network branch considering patientspecific clinical data, such as demographic information, medical history, and laboratory test results, was taken into account and concatenated with the imaging features for a holistic analysis. DeepRetina achieved 95% sensitivity, 98% specificity for early DR detection, and a 0.92 area under the curve (AUC) for 5-year progression prediction, outperforming existing methods.
Volume: 15
Issue: 1
Page: 152-160
Publish at: 2026-03-01
1 2 3
...
2062 2063
Show 86 of 2063

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