JUTI: Jurnal Ilmiah Teknologi Informasi
Not a member yet
407 research outputs found
Sort by
Mandibular Image Segmentation and 3d Reconstruction using U-Net Model
Full text linkPenelitian ini bertujuan untuk meningkatkan presisi dan efisiensi dalam segmentasi citra mandibula dan rekonstruksi 3D menggunakan model U-Net. Segmentasi otomatis dengan U-Net menangani tantangan metode manual yang memakan waktu. Struktur Encoder-Decoder pada U-Net memungkinkan pembelajaran fitur citra medis yang kompleks dengan akurasi tinggi, menghasilkan segmentasi yang konsisten dan presisi. Hasil penelitian menunjukkan bahwa Res U-Net mencapai performa segmentasi yang unggul dengan Dice Similarity Coefficient (DSC) sebesar 95,37%, meskipun memerlukan waktu komputasi yang lebih lama. Sementara itu, U-Net standar menawarkan efisiensi komputasi yang lebih tinggi dan cocok untuk aplikasi real-time meskipun akurasinya sedikit lebih rendah. Integrasi segmentasi dengan rekonstruksi 3D meningkatkan visualisasi anatomi mandibula, memperbaiki efektivitas perencanaan bedah, serta menyediakan alat simulasi interaktif untuk perawatan personal dan pelatihan profesional. Penggunaan standar DICOM memfasilitasi aksesibilitas antar perangkat medis, mendukung interoperabilitas sistem perawatan kesehatan. Studi ini menyimpulkan bahwa Res U-Net optimal untuk kebutuhan presisi tinggi, sedangkan U-Net lebih cocok untuk aplikasi dengan pemrosesan cepat. Temuan ini diharapkan dapat memajukan teknologi segmentasi dan visualisasi medis yang andal dan efektif dalam praktik klinis
Enhancing Face Detection Performance In 360-Degree Video Using Yolov8 with Equirectangular Augmentation Techniques
Full text linkThis study aims to enhance face detection performance in 360-degree videos by utilizing advanced image augmentation techniques with the YOLOv8 algorithm, which is effective for real-time object detection. Acknowledging the unique challenges posed by equirectangular projection, this research introduces a novel equirectangular augmentation method specifically designed for this medium. Our findings demonstrate a remarkable 1.346% improvement in detection accuracy in Equirectangular Projection (ERP) settings compared to default YOLOv8 augmentation strategies. This significant enhancement not only addresses the geometric distortions inherent in panoramic video formats but also emphasizes the critical need for tailored augmentation approaches to improve face detection in complex environments. By showcasing the effectiveness of these customized methods, this research contributes to the growing field of deep learning applications for immersive video technologies, with implications for sectors like security, virtual reality, and interactive media. Ultimately, this work highlights the potential of innovative augmentation techniques to ensure robust face detection in challenging visual contexts
Topic Modeling for Constructing Learning Profiles Using LDA and Coherence Evaluation
No full textUnderstanding individual learning patterns is important for supporting effective learning strategies in the digital education ecosystem. This study proposes a topic modeling approach using the Latent Dirichlet Allocation (LDA) algorithm to form learning profiles based on student interaction data from EdNet-KT1. The dataset includes 153,824 interactions with 11,613 questions, which were converted into semantic tag-based pseudotexts. Modeling was performed with 20 topics, which were selected as a compromise between semantic quality (coherence score 0.6688) and model readability, although the highest coherence score appeared with a larger number of topics. Each question is linked to a dominant topic, and student accuracy is calculated to form a student-topic performance matrix. The results of the analysis show that 66% of students mastered more than five topics, reflecting a broad range of knowledge. Visualization with heat maps, radar charts, and line charts provides a detailed overview of each individual\u27s strengths and weaknesses. Segmentation was performed using the K-Means algorithm and produced four clusters based on student performance distribution. Adaptive learning recommendations are compiled based on an accuracy threshold of < 0.5 and a number of interactions > 10. Topics_13, topics_10, and topics_12 were identified as the most challenging topics. The results of this study indicate the potential of LDA-based approaches and clustering as analytical tools for shaping more personalized and contextual learning systems. Further research could explore sequential modeling and experimental validation of the effectiveness of recommendation
Evaluation of Synthetic Data Effectiveness using Generative Adversarial Networks (GAN) in Improving Javanese Script Recognition on Ancient Manuscript
Full text linkThe imbalance of Javanese script data in ancient manuscript recognition poses a challenge due to the limited availability of data. A potential approach to addressing this issue is the use of Generative Adversarial Networks (GAN). This study evaluates the effectiveness of synthetic data generated using Enhanced Balancing GAN (EBGAN) in mitigating data imbalance. Various evaluation scenarios are conducted, including: (i) assessing the impact of syn-thetic data as augmentation, (ii) evaluating the sufficiency of synthetic data for recognition models, (iii) analyzing minority class oversampling with different selection strategies, and (iv) evaluating model generalization through cross-validation. Quantitative analysis of the generated synthetic data, based on Fréchet Inception Distance (FID) and Structural Similarity Index (SSIM), as well as visual assessment, indicates that the quality of synthetic data closely resembles real data. Additionally, experimental results demonstrate that combining real and synthetic data improves accuracy, precision, recall, and F1-score. The oversampling strategy for synthetic data has proven effective in meeting the data sufficiency requirements for training recognition models. Meanwhile, selecting minority classes and determining threshold values based on percentage, distribution, and model performance in oversampling can serve as guidelines for enhancing script recognition performance. Compared to previous methods, the use of EBGAN has been shown to produce more diverse synthetic data with better visual quality. However, further research is still needed to optimize GAN performance in supporting script recognition
Exploring The Effectiveness of In-Context Methods in Human-Aligned Large Language Models Across Languages
Full text linkMost of past studies about in-context methods like in-context learning (ICL), cross-lingual ICL (X-ICL), and in-context alignment (ICA) come from older, unaligned large language models (LLMs). However, modern human-aligned LLMs are different; they come with chat-style prompt templates, are extensively human-aligned, and cover many more languages. We re-examined these in-context techniques using two recent, human-aligned multilingual LLMs. Our study covered 20 languages from seven different language families, representing high, mid, and low-resource levels. We tested how well these methods generalized using two tasks: topic classification (SIB-200) and machine reading comprehension (Belebele). We found that utilizing prompt templates significantly improves the performance of both ICL and X-ICL. Furthermore, ICA proves particularly effective for mid- and low-resource languages, boosting their f1-score by up to 6.1%. For X-ICL, choosing a source language that is linguistically similar to the target language, rather than defaulting to English, can lead to substantial gains, with improvements reaching up to 21.98%. Semantically similar ICL examples continue to be highly relevant for human-aligned LLMs, providing up to a 31.42% advantage over static examples. However, this gain decreases when using machine translation model to translate query from target language. These results collectively suggest that while modern human-aligned LLMs definitely benefit from in-context information, the extent of these gains is highly dependent on careful prompt design, the language\u27s resource level, language pairing, and the overall complexity of the task
Mixed-Integer Linear Programming for Optimal Operation of Integrated Electricity and Natural Gas System Considering Take or Pay Agreements
Full text linkThis paper is proposed to demonstrate the implementation of Mixed-Integer Linear Programming (MILP) for solving the optimal operation of the Integrated Electricity and Natural Gas System (IENGS). The MILP is used to realize an economical and reliable power electricity system based on Dynamic Optimal Power and Gas Flow (DOPGF) considering Take or Pay (TOP) agreements for natural gas. This method is simulated on the integrated 6-bus electricity and 6-node natural gas systems. By using MILP, the best costs for optimal operation of IENGS are obtained in three scenarios. The superiority of the MILP is validated by suppressing the increasing best cost for optimal operation to be below 10%. In the first case, the best cost is 748,399.30 to 791,833.04 with the TOP agreement in all of the generators, which is 7.67% higher than the first scenario. In addition, the MILP can perform the DOPGF for IENGS without violating the problem constraints regarding the load demand fulfillment and power system limitations in both coal-fired and gas-fired generators.
Survey on Risks Cyber Security in Edge Computing for The Internet of Things Understanding Cyber Attacks Threats and Mitigation
Full text linkDalam era pesatnya perkembangan teknologi, penggunaan IoT terus meningkat, terutama dalam konteks edge computing. Makalah survei ini secara teliti menjelajahi tantangan keamanan yang muncul dalam implementasi IoT pada edge computing. Fokus utama penelitian ini adalah potensi serangan dan ancaman siber yang dapat mempengaruhi keamanan sistem. Melalui metode survei literatur, makalah ini mengidentifikasi risiko keamanan siber yang mungkin timbul dalam lingkungan IoT di edge computing. Pendekatan metodologi penelitian digunakan untuk mengklasifikasikan serangan berdasarkan dampaknya pada infrastruktur, layanan, dan komunikasi. Keempat dimensi klasifikasi, yaitu Network Bandwidth Consumption Attacks, System Resources Consumption Attacks, Threats to Service Availability, dan Threats to Communication, memberikan dasar untuk memahami dan mengatasi risiko keamanan. Makalah ini diharapkan memberikan landasan pemahaman yang kokoh tentang keamanan pada IoT dalam edge computing, serta kontribusi untuk pengembangan strategi keamanan yang efektif. Dengan fokus pada pemahaman mendalam tentang risiko keamanan, makalah ini mendorong pengembangan solusi keamanan yang adaptif di masa depan untuk mengatasi tantangan keamanan yang berkembang seiring dengan pesatnya adopsi teknologi IoT di edge computing
EXPLORING CONSUMER RESPONSE TO TEXT-BASED CHATBOTS IN F-COMMERCE: A QUALITATIVE STUDY ON BANGLADESHI SME’S
Full text linkThis qualitative study examines the consumer response to text-based chat bots in F-commerce, specifically in the context of Bangladeshi SMEs. The study aims to explore the benefits and challenges of using chat bots in F-commerce and identify the factors that influence consumer response to chat bots. The study uses semi-structured interviews to collect data from 15 Bangladeshi consumers who have experience using chat bots in F-commerce. The findings suggest that chat bots can improve customer service, save time and effort, and provide convenience for consumers, but they also face challenges such as technical issues, language barriers, and privacy concerns. The study also identifies several factors that influence consumer response to chat bots, including perceived usefulness, perceived ease of use, trust, familiarity, and personalization. The study concludes by discussing the practical implications of the findings for SMEs in Bangladesh and suggesting directions for future research
RISK MANAGEMENT ANALYSIS OF E-KOHORTKIA APPLICATION USING ISO 31000 FRAMEWORK IN SOUTH CENTRAL TIMOR DISTRICT HEALTH OFFICE
Full text linkThe E-KohortKIA application is a web-based application launched by the Indonesian Ministry of Health for use in Puskesmas and Health Offices throughout Indonesia as a solution to various problems caused by manual use of the KIA Cohort. This application can be accessed via computer or smartphone and is currently being implemented by the South Central Timor District Health Service for Maternal and Child Health (KIA) services. In using this application, various possible risks can occur that can disrupt application performance, therefore it is necessary to carry out a risk management analysis. The aim of this research is to determine various possible risks that could occur in implementing the E-KohortKIA application and to carry out risk treatment for these possible risks. This research uses qualitative methods by collecting data through interviews and observations, as well as data processing and risk management analysis using the ISO 31000 framework including the risk assessment stage and risk treatment stage. The results of this research found 23 possible risks, most of which were due to system and infrastructure factors. These possible risks include 13 possible low level risks, 8 possible medium level risks, and 2 possible high level risks. There are 2 possible risks that have a high and maximum level of risk severity so that they have the potential to disrupt or inhibit or even stop application performance. This research also provides recommendations for risk treatment proposals for various possible risks and can be used by users to maintain application performance
OVERSAMPLING HYBRID METHOD FOR HANDLING MULTI-LABEL IMBALANCED
Full text linkData and information continue to increase along with the development of digital technology. Data availability is becoming increasingly numerous and complex. The existence of unbalanced data causes classification errors due to the dominance of majority-class data over the minority class. Not only limited to the binary class, but data imbalance is also often encountered in multi-label data, which become increasingly important in recent years due to its vast application scope. However, the problem of class imbalance has been a characteristic of many complex multi-label datasets, making it the focus of this research. Handling unbalanced multi-label data still has a lot of potential for development. One approach, Synthetic Oversampling of Multi-Label Data Based on Local Label Distribution (MLSOL) and Integrating Unsupervised Clustering and Label-specific Oversampling to Tackle Imbalanced Multi-Label Data (UCLSO), has been developed. UCLSO\u27s attention only focuses on the majority class, which can lead to data imbalance and excessive overfitting. Although effective in preventing majority class domination, this approach cannot overcome the lack of variation within the minority class. By contrast, MLSOL focuses on minority classes, allowing for variations in multi-label data and significantly improving classification performance. This research aims to overcome the problem of data imbalance by combining the MLSOL and UCLSO oversampling methods. Combining these two approaches is expected to exploit the strengths and reduce the weaknesses of each, resulting in significant performance improvements. The trial results show that the hybrid oversampling method produces the highest value on biological data with an F-1 score of 88%. Meanwhile, the single oversampling methods UCLSO and MLSOL on biological data produce an F-1 score of 67% and 62%, respectively