|
|
 |
Hybrid Event
|
| Papers |
T. Alajbeg, J. Ećimović, M. Sokele (Zagreb University of Applied Sciences, Zagreb, Croatia), M. Mačković (Fiditas-explosion safety solutions d.o.o., Zagreb, Croatia)
Determination the Temperature Class for Explosion Proof Equipment 
The aim of the conformity assessment procedure is to prove that equipment intended to operate in the presence of flammable or explosive atmospheres will not ignite such atmospheres. Therefore, for such equipment, it is necessary to correctly determine the temperature class considering the environment in which it will operate. To determine the temperature class, measurements are carried out using appropriate testing equipment in accordance with the requirements of standards and the relevant Laboratory procedure of the International Electrotechnical Commission for the conformity assessment for Electrotechnical Equipment and Components (IECEE). The temperature class is mandatory information that has to be included in the product's technical documentation - component datasheet, or stated on the equipment housing. This paper presents the measuring process and the data obtained and processed by a computer for several electrotechnical explosion-proof products. Based on the obtained results, the temperature class is determined. The paper was motivated and created based on a master’s thesis at the graduate study of Electrical Engineering at the Zagreb University of Applied Sciences.
|
E. Starić, N. Tanković (Juraj Dobrila University of Pula, Pula, Croatia)
Preliminary Study on the Difference of Web Services Load Balancing Performance in Theory and Practice
Web service load balancing is a process of distributing workload over multiple servers to support and ensure high availability and better performance in modern applications. The purpose of load balancing is to reduce the number of tasks each server processes to make them more efficient, increase performance, and reduce latency. The theoretical aspect of load balancing can be studied through queueing theory - a branch of mathematics that studies how lines form and how they function to predict queue length and time to process a task in a queue. The results can then be applied to achieve better load balancing. However, queueing theory does have some limitations; that is, it relies on some assumptions (e.g., arrival rate of tasks, and their complexity), which can lead to different behavior in practice than what was predicted. This paper aims to compare the expected results of several web service load balancing algorithms got through queueing theory and the performance of the same web service load balancing algorithms in practice to see if there are any differences, and if there are, how big they are.
|
E. Mirković, V. Kirinčić, R. Prenc, M. Rojnić (University of Rijeka, Faculty of Engineering, Rijeka, Croatia)
Decarbonization through Hybrid Energy Systems
Decarbonization through hybrid energy systems represents a key strategy for reducing carbon emissions. Hybrid energy systems play a crucial role in the advancement of energy infrastructure. This paper presents an analysis of a household hybrid energy system comprising a photovoltaic (PV) system, battery energy storage system (BESS), inverter, and grid connection. The system design is developed using several software tools, followed by simulation. The simulation results are analyzed, revealing positive outcomes for the decarbonization of the energy sector.
|
M. Tomašević, V. Zuppa Bakša, T. Tolnauer Ackermann (Zagreb University of Applied Sciences, Zagreb, Croatia)
The Role of Hydroelectric Power Plants in Power System Stabilization
This paper analyzes the power system, its main components and the factors affecting its stability, such as sudden load changes, faults, and disturbances. Furthermore, it explains the operating principles of hydroelectric power plants and provides a detailed description of their components. Special attention is given to the role of hydroelectric power plants in stabilizing the power system, particularly focusing on storage hydroelectric power plants as key elements for frequency and power exchange regulation. Finally, the advantages and disadvantages of hydroelectric power plants compared to other energy sources are discussed.
|
K. Bešenić, V. Zuppa Bakša, A. Bednjanec (Zagreb University of Applied Sciences, Zagreb, Croatia)
Economic Aspects of Electricity Production in Wind Power Plants
The goal of this paper is to examine key economic aspects related to onshore wind power plants and their evolution over the years—such as the Levelized Cost of Electricity (LCOE), total costs associated with wind power plant construction in various countries, operational and maintenance costs incurred throughout the lifecycle of wind turbines, and the potential for job creation within the wind energy sector. Finally, the paper will present a conclusion regarding the economic viability of wind power plants.
|
I. Črnjak, M. Sokele, G. Vujisić (Tehničko veleučilište u Zagrebu, Zagreb, Croatia)
Development of the device and mobile application for collecting, analyzing and displaying vehicle diagnostic data
The paper presents the hardware and software development of the device, which enables the collection of vehicle diagnostic data as well as a mobile application for visualization of collected data. In addition, the application adds the corresponding geographic coordinates of the vehicle and the current time stamp to the collected diagnostic data. The emphasis of the paper is on data collection using the CAN interface and data processing based on a microcontroller. The mobile application operates under an open-source framework and can be part of the educational process for further application development and for conducting hardware-oriented laboratory exercises. Testing the functionality of the device and the mobile application was confirmed on driving routes near the City of Zagreb.
|
A. Vedriš, B. Tomičić (Zagreb University of Applied Sciences, Zagreb, Croatia), D. Sente (Zagreb University of Applied Sciences, Zagreb, Croatia)
Modeling of Permanent Magnet Generators Using the Finite Element Method (FEM)
The paper provides an overview of the types of synchronous machines with permanent magnets. An energy conversion system using a high-speed turbine direct coupled with permanent magnet synchronous generator is described. Using the example of a permanent magnet generator, a model of the generator was created in the MagNet software tool using FEM. Due to the high speed of rotor, and thus the high frequency of electric parameters, special attention was given to the properties of materials embedded in the machine e.d. the selection of magnets and magnetic sheets. Simulation of operation at no load and at full load of the generator were performed. Results of both of those tests gave the insight into the electromagnetic conditions inside the generator and they came up as satisfactory considering power and efficiency at the nominal speed.
|
S. Kroflin, T. Jakopec (Filozofski fakultet Osijek, Osijek, Croatia)
An Analysis of the Performance of .hr Domain Homepages Using the Google Lighthouse Tool
The paper analyzes the performance of homepages within the .hr domain using the Google Lighthouse tool. The analysis includes 111,117 homepages with data collected on key performance metrics: First Contentful Paint (FCP), Largest Contentful Paint (LCP), Speed Index, Total Blocking Time (TBT), and Cumulative Layout Shift (CLS). In addition to these metrics, data related to console errors, the number of network requests (HTTPS), page sizes, and the use of the HTTP(S) protocol were analyzed. The data were collected using a custom software solution and stored in a MariaDB relational database. The analysis was performed using SQL queries. The query results indicate frequent problems such as console errors, unoptimized image content, and excessive use of resources that prolong website loading times. The results are compared with previous studies, i.e., their results and data, to gain a better understanding of the performance of .hr homepages compared to other pages. The discussion and conclusion offer suggestions for website optimization regarding the metrics mentioned and provide useful insights for future research into the performance of homepages within the .hr domain.
|
A. Hasanović, D. Pozderac, E. Cogo, E. Cogo, I. Bešić (University of Sarajevo, Faculty of Electrical Engineering, Sarajevo, Bosnia and Herzegovina)
Comparison of MSP-EXP432P401R Program Performances by Using Different Development Environments
Texas Instruments development kits have a wide application in practical and scientific experiments due to their small size, processing power, available booster packs, and compatibility with different environments. The most popular integrated development environments for programming these development kits are Energia and Code Composer Studio. Unfortunately, there are no existing studies that compare the benefits and drawbacks of these environments and their performances. Conversely, the performances of the FreeRTOS environment are well-explored, making it a suitable baseline for embedded systems execution. In this paper, we performed the experimental evaluation of the performance of Texas Instruments MSP-EXP432P401R when using Energia, Code Composer Studio, and FreeRTOS for program execution. Three different sorting algorithms (bubble sort, radix sort, merge sort) and three different search algorithms (binary search, random search, linear search) were used for this purpose. The results show that Energia sorting algorithms outperform other environments with a maximum of 400 elements. On the other hand, FreeRTOS search algorithms far outperform other environments with a maximum of 255,000 elements (whereas this maximum was 10,000 elements for other environments). Code Composer Studio resulted in the largest processing time, which indicates that the low-level registry editing performed in this environment leads to significant performance issues.
|
| Break |
| Papers |
D. Švraka, A. Ali Al Zayat, E. Cogo, E. Cogo, I. Bešić (University of Sarajevo, Faculty of Electrical Engineering, Sarajevo, Bosnia and Herzegovina)
Evaluation of the Usability of Analog Joystick Movement in MSP-EXP432P401R Snake Game
Computer games can be used not only for entertainment but also for education. Embedded systems can be used to improve the gamified learning process by making the interaction with intended users more interesting. Texas Instruments development kits with booster plug-in modules can improve the outcomes of gamified learning. However, there is a lack of studies that explore the benefits and drawbacks of different input methods for gamified learning purposes. In this paper, a snake game was developed on the Texas Instruments MSP-EXP432P401R development kit that uses the analog joystick of the BOSTXL-EDUMKII plug-in module for controlling the snake. An experimental usability study was conducted on 61 3rd year students, comparing the analog joystick to the computer keyboard, computer mouse, and mobile touchscreen input methods. The achieved results showed that the majority of students preferred the original computer keyboard input method and that more than half of the participants preferred the 90-degree rotation of the snake compared to the 360-degree analog joystick. However, the analog joystick improved the gaming experience by 63.6\%, and many students made positive comments about its usability in general, indicating that its application for gamified learning may be possible for other types of games.
|
E. Smajić, E. Cogo, E. Cogo, I. Bešić (University of Sarajevo, Faculty of Electrical Engineering, Sarajevo, Bosnia and Herzegovina)
IoT-based Braille Alphabet Input and Translator System
The blind and visually impaired group cannot use most of the cutting-edge technology that usually conveys information visually through different kinds of displays. Different solutions can help overcome this obstacle, such as the usage of sound output and tactile displays that use the Braille alphabet composed of mechanically raised dots. However, there is a considerable amount of visually impaired persons who cannot read Braille and an ever larger amount of persons without visual impairment. This paper presents an IoT-based system that uses the Arduino Uno WiFi development board for reading Braille input from a 4x4 push button matrix, two letters at a time. The system uses the 32x8 matrix display to show the translated basic alphabet output that can be read by sighted users or to show the Braille alphabet output. It offers a quick way for the visually impaired to convey information to sighted people by typing Braille input with both hands simultaneously. The proposed system will be used to educate sighted individuals about the Braille alphabet and help reduce their learning time. It can also be used as a quick translator of Braille for sighted individuals who wish to read written Braille documents.
|
D. Radošević, I. Lacković (University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, Croatia)
Contribution to Human Respiratory System Modelling Using Electrical Networks
This study presents a comprehensive model of the human respiratory system using electrical circuit analogies to explore its acoustic and mechanical characteristics. The model segments the respiratory tract into uniform components, each represented by T-circuits comprising resistance, capacitance, inductance, and conductance. The study investigates the effects of structural changes, like variations in airway length and cross-sectional area, on sound signal features, including the first formant frequency and intensity. Graph theory was employed to model branching structures in the subglottic region, enhancing the model’s anatomical accuracy. The study also examines sound propagation through the pulmonary parenchyma, analyzing attenuation and acceleration as functions of tissue and gas proportions. Practical applications include non-invasive respiratory monitoring and diagnosing conditions like apnea and pulmonary edema. By parameterizing the model, diverse structures of the respiratory system have been simulated. This work advances the understanding of respiratory acoustics, supporting the development of diagnostic tools for detecting and managing respiratory disorders.
|
V. Imrović, A. Bednjanec, V. Zuppa Bakša (Zagreb University of Applied Sciences, Zagreb, Croatia)
Environmental Management in Electrotechnical Products Manufacturing Process
Electronic waste (e-waste) represents an escalating global environmental issue due to the presence of hazardous substances including lead, mercury, and cadmium. The ISO 14001 standard provides a systematic framework for environmental management, allowing organisations to minimise their ecological impact via resource optimisation, waste reduction, and sustainable process management. The implementation of these norms in electrotechnical manufacturing processes contributes to sustainable development by increasing product lifespans, facilitating repairs, and decreasing energy consumption. The critical phases of the product life cycle - pre-production, production, and post-production - are aligned with the standard's criteria to mitigate environmental and health risks. This article emphasises critical elements such material recycling, reduction of carbon footprint, and the implementation of sustainable technologies. Ultimately, it underscores the significance of education, the enhancement of recycling capabilities, and the promotion of ecologically sustainable activities as vital instruments for addressing the issues of e-waste and its effects on ecosystems.
|
R. Turčinović, G. Mikić (Sveučilište u Rijeci, Fakultet informatike i digitalnih tehnologija, Rijeka, Croatia), D. Pitoski (Sveučilište u Rijeci, Centar za umjetnu inteligenciju i kibernetičku sigurnost, Rijeka, Croatia), S. Beliga, A. Meštrović (Sveučilište u Rijeci, Fakultet informatike i digitalnih tehnologija, Rijeka, Croatia)
Analyzing Social Networks of Sport Preferences for Personalized Recommendations
Social network analysis (SNA) has emerged as a powerful tool for enhancing recommender systems, particularly in domains where user preferences and interactions play a significant role. This paper explores the application of SNA techniques to a network constructed from user sport preferences. We conducted a multi-level analysis of the network, examining its global, meso, and local structures to uncover patterns and relationships. Key network measures, such as centrality and community structure, were identified as critical for generating effective user recommendations.
Building on these findings, we propose a novel approach that combines community detection with other network metrics to recommend new connections within the network. Our method is designed to foster meaningful interactions by suggesting training partners based on shared interests and complementary roles within the network. The results demonstrate the potential of integrating SNA with recommendation algorithms to enhance user experiences and promote engagement in sport-related activities.
|
R. Katalinić, M. Rabađija, A. Meštrović, S. Beliga (Sveučilište u Rijeci, Fakultet informatike i digitalnih tehnologija, Rijeka, Croatia)
Mreže supojavljivanja entiteta u medijskim tekstovima za analizu političkih kampanja
U ovom radu istražene su mogućnosti analize medijskih tekstova povezanih s političkim kampanjama pristupom koji kombinira metode računalne analize prirodnoga jezika i analize društvenih mreža. Konkretno, s ciljem definiranja postupka za proučavanje diskursa i promjena tijekom političke kampanje iz korpusa medijskih tekstova ekstrahiraju se politički akteri. Potom se konstruiraju mreže supojavljivanja političkih entiteta koje se analiziraju na globalnoj, središnjoj i lokalnoj razini. Primjena predloženog pristupa prikazana je kroz analizu medijskih tekstova koji su objavljeni tijekom predizborne kampanje za parlamentarne izbore 2024. godine u Republici Hrvatskoj. Rezultati potvrđuju da predloženi pristup pruža uvid u obrasce pozicioniranja i povezivanja političkih subjekata u medijskom prostoru te otkriva strukturalne promjene što doprinosi razumijevanju dinamike interakcija u političkom diskursu tijekom kampanje.
|
S. Jakjimovski, A. Vlahovski, M. Stojcheva, A. Dedinec ( Faculty of Computer Science and Engineering Ss. Cyril and Methodius University, Skopje, Macedonia), J. Prodanova (Macedonian Academy of Sciences and Arts, Skopje, Macedonia), F. Liébana Cabanillas (Faculty of Economics and Business Administration University of Granada, Spain, Granada, Spain), L. Kocarev (Macedonian Academy of Sciences and Arts, Skopje, Macedonia)
Sentiment Analysis of Pro-Environmental Behavior using Twitter data: A Comparative Study of Waste Management in Spain and Macedonia
This research examines pro-environmental behavior through sentiment analysis of Twitter data, focusing on waste management discussions in Spain and Macedonia from May 2024 to the present. We employ BERT-based models—Multilingual BERT and BERTweet—to analyze sentiment and perform topic modeling using BERTopic. For Spanish tweets, native-language analysis with Multilingual BERT outperforms BERTweet when tested on translated data, revealing a predominantly positive sentiment (80%). In contrast, Macedonian tweets, translated into English and evaluated on both models, show consistent performance but reflect an opposite trend, with 70% of tweets expressing negative sentiment. Sentiment analysis categorizes tweets into positive, neutral, and negative classes, enabling a comparative evaluation of model effectiveness across translations and languages. Additionally, BERTopic identifies prominent themes both across the entire dataset and within individual sentiment categories, offering insights into public attitudes and discussions. These findings highlight regional differences in public perceptions of waste management and the value of multilingual approaches to sentiment and topic analysis. This research contributes to understanding environmental communication on social media and the effectiveness of language models in capturing diverse regional behaviors.
Keywords: Sentiment Analysis, Waste Management, Twitter Data, Multilingual BERT, BERTweet, BERTopic, Pro-Environmental Behavior
|
V. Pavlović, K. Širić, B. Mihaljević (RIT Croatia, Zagreb, Croatia), A. Radovan (Algebra University College, Zagreb, Croatia)
Optimizing Java Performance with the Vector API for Data-Parallel Computations
The Java Vector API efficiently optimizes data-parallel operations using SIMD (Single Instruction, Multiple Data), making it particularly useful for applications that rely on vector-based numerical computations, such as simulations, machine learning, and data analytics. This paper analyzes the Vector API's design, supported features, and potential use in performance-critical domains such as game engines, data analytics, and scientific simulations.
We analyzed the Vector API with practical test cases to measure its performance, focusing on features such as supported data types, operational flexibility, and the way older CPUs without appropriate SIMD support affect its efficiency. Our preliminary results demonstrate notable improvements in the performance of parallelizable tasks while also identifying limitations, including reliance on hardware SIMD support and reduced efficiency when fallback mechanisms are employed.
The Vector API demonstrates how Java can leverage hardware-accelerated computations, allowing software engineers to achieve significant performance gains without requiring the use of low-level programming languages. The findings suggest that the Vector API effectively bridges the gap between hardware-accelerated computations and Java’s high-level programming model, providing a viable path for optimizing performance in specific use cases without requiring low-level programming expertise.
|
| Papers |
M. Šimatović, H. Markulin, D. Beronić, B. Mihaljević (RIT Croatia, Zagreb, Croatia)
Evaluating Memory Management and Garbage Collection Algorithms with Virtual Threads in High-Concurrency Java Applications
Garbage collection (GC) is a critical memory management mechanism within the Java Virtual Machine (JVM) responsible for automating memory allocation and reclamation. Its performance affects the computational efficiency, scalability, and responsiveness of Java applications, particularly in high-concurrency environments with demanding resource utilization patterns. Optimizing memory management of highly concurrent applications can enhance performance by reducing latency, improving throughput, and ensuring efficient resource utilization.
With the introduction of a JVM-based lightweight threading model named Virtual Threads, the dynamics between concurrency constructs and GC algorithms have gained renewed importance. Virtual Threads enable the creation of millions of threads, imposing unique demands on memory management systems.
This paper presents a comparative analysis of several modern GC algorithms, focusing on their performance in Virtual Thread-enabled environments. With a set of benchmarks simulating real-world workloads, we evaluated key performance metrics, including CPU usage, heap efficiency, latency, and throughput, thus encompassing both computationally intensive and I/O-bound tasks that reflect
scenarios encountered in modern applications.
The results reveal the trade-offs in GC performance, particularly handling the rapid thread creation, memory allocation, and high task-switching rates characteristic of Virtual Threads. Our findings offer insights into optimizing memory management and performance of contemporary high-concurrency Java applications using Virtual Threads.
|
F. Alaber, B. Novoselnik (University of Zagreb Faculty of Electrical Engineering and Computing, Zagreb, Croatia)
IoT-Driven Weather Station for Comprehensive Data Collection
This technical paper focuses on the development of a simple weather station that is easy to replicate with minimal costs. The current version of the station measures temperature, humidity, and precipitation, with the collected data being stored locally. The motivation for the project stems from the fact that the Zagreb area currently has only four weather stations, which, although technologically advanced, are limited in spatial coverage. By utilizing affordable microcontrollers, sensors, and 3D additive technologies, it is possible to produce a larger number of localized stations. Their distribution across Zagreb would enable more detailed measurements, contributing to more accurate weather forecasts, early warning systems, and scientific models. The current research focus is on the prototype station and the comparison of its data with those from reference meteorological stations. Future development aims to enhance system autonomy by integrating solar panels and batteries, eliminating dependence on local Wi-Fi networks, and creating infrastructure for data transmission to a central server from a network of stations distributed throughout the city.
|
E. Gill (University "G. d'Annunzio" Chieti-Pescara, Italy, Pescara, Italy), M. Mastromatteo (Management HSQE Field & Safety Operations, Italferr S.p.A, Firenze, Italy), L. Cangelmi, P. Cellini (University "G. d'Annunzio" Chieti-Pescara, Pescara, Italy), D. Cardone, A. Amelio (University "G. d'Annunzio" Chieti-Pescara, Italy, Pescara, Italy)
A Comparative Analysis of Artificial Intelligence Methods for Air Quality Prediction on Construction Sites
This study focuses on predicting air pollutants on construction sites, which is an essential aspect for preserving the health of workers and people who live nearby. We carried out through data pre-processing, handling missing values and transforming categorical variables. The focus is forecasting different key air pollutants like PM2.5, PM10, SO2, NO, CO2, and O3. Hence, to achieve this, we compare different statistical, machine learning and deep learning approaches. The novelties of the proposed work from the previous works are: (i) the prediction of multiple pollutants, (ii) the use of multiple predictive models, and (iii) a larger prediction window of 12 hours. We compare the models by computing the Root Mean Squared Error and the R2 to assess their performances. This study provides a comparative analysis of well-known models in the literature for predicting air quality in construction sites. In conclusion, the findings show that training LSTM models can significantly enhance air pollution predictions, providing valuable insights for improving environmental monitoring and forecasting accuracy.
|
G. Perković, A. Drobnjak (Faculty of Electrical Engineering and Computing University of Zagreb, Zagreb, Croatia), I. Terzić (LAFCO TECH, Zagreb, Croatia), I. Botički (Faculty of Electrical Engineering and Computing University of Zagreb, Zagreb, Croatia)
Reaching Artificial General Intelligence (AGI): Challenges and Opportunities
Artificial General Intelligence (AGI) presents a groundbreaking milestone in artificial intelligence (AI) development, and could match, or even surpass human intellectual capabilities. This paper investigates AGI development and discusses the balance between the risks and benefits to our society associated with the AGI use. The most important risks include social and economic instability, user privacy and cybersecurity threats. On the other hand, the benefits could include significant advances across the branches of science. The paper examines the architectural challenges, draws the comparison between AGI and other AI models, and discusses ethical dilemmas in AGI use. In conclusion, the impact of risks and opportunities is discussed, giving a perspective on the future of AGI and its impact on humanity.
|
J. Kolarec, J. Petrović (University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, Croatia)
An Overview of Passive Digital Image Tampering Detection Methods
Forgery detection and digital image authentication are important tasks in digital image forensics. Various methods for forgery detection and digital image authentication have been proposed and evaluated in the literature, typically based on neural networks and deep learning, image processing algorithms, mathematical techniques, and image blocks and keypoints detection. Passive image forgery detection methods, unlike active methods, do not require any previously embedded information to establish the authenticity of an image. This paper provides an overview of passive digital image tampering detection methods and highlights strengths, limitations, and potential areas for improvement of each proposed approach, with particular emphasis on techniques used to identify copy-move forgeries.
|
J. Bernetić, M. Ašenbrener Katić, S. Čandrlić (Sveučilište u Rijeci, Fakultet informatike i digitalnih tehnologija, Rijeka, Croatia)
Developing a Child-Friendly Chatbot Using GPT and OpenAI API: A Case Study
The development of modern technology has led to significant advancements in various aspects of society, including the field of artificial intelligence (AI). One of the key breakthroughs in this area is the development of large language models (LLMs), based on massive amounts of textual data. These models enable computers to comprehend and generate human language with an exceptional level of coherence and accuracy.
Examples of popular language models include OpenAI's ChatGPT systems, which have found wide-ranging applications across industries—from automating customer support to creating creative content. Their ability to generate contextually appropriate responses makes them suitable for diverse scenarios.
This paper provides a detailed description of the process of creating a chatbot based on the GPT-3.5 model. It outlines the steps involved in building the chatbot and connecting the model via OpenAI's API, using prompt engineering techniques. The outcome is a customized chatbot capable of generating responses suitable for children, with a special emphasis on comprehensibility and an adjusted linguistic style.
|
N. ITHAL, L. Suryanarayana Sharma, S. M, K. Kayartaya, R. Seethur (PES University, Bangalore, India)
Efficient Implementation of Error Sampling Block for Homomorphic Encryption Using High-Level Synthesis
In the era of secure data processing, Homomorphic Encryption (HE) has emerged as a groundbreaking technique, enabling computations on encrypted data without decryption, thereby preserving data privacy. The CKKS (Cheon-Kim-Kim-Song) scheme, renowned for supporting approximate arithmetic, is particularly suitable for applications like machine learning and data analytics. A vital element in HE is the error sampling block, which impacts both the performance and security of the encryption scheme. This paper focuses on the high-level synthesis (HLS)-based implementation of an error sampling block tailored for the CKKS scheme. By leveraging Cadence Stratus HLS with the GPDK 45nm library and SystemC coding, we demonstrate the potential of HLS in streamlining hardware design for cryptographic applications. Our implementation incorporates a low-area Pseudo-Random Number Generator (PRNG) and optimized sampling methods, achieving a 41.16% reduction in power and a 7.13% reduction in area. These results highlight the benefits of HLS, such as accelerated design cycles, reduced complexity, and suitability for resource-constrained environments.
|
S. Trojer (University of Ljubljana, Ljubljana, Slovenia), V. Bhateja (Veer Bahadur Singh Purvanchal University, Jaunpur, India), P. Peer, Ž. Emeršič (University of Ljubljana, Ljubljana, Slovenia)
DeepCombine: Multimodal Person Recognition
The paper explores multimodal biometric recognition by integrating face, iris, and fingerprint data to enhance person identification accuracy and robustness. We compare the effectiveness of late fusion with traditional unimodal approaches, demonstrating that modality fusion significantly improves classification performance. A combined dataset was created by merging CelebFaces, Casia-Iris-Thousand, and Casia Fingerprint datasets, with preprocessing steps to address demographic and sample imbalances. We evaluate feature extraction models, including EfficientNet and FaceNet, and employ machine learning classifiers such as k-Nearest Neighbors (kNN) and Random Forest (RF). Our experiments show that kNN, combined with Principal Component Analysis (PCA) for dimensionality reduction, achieves perfect classification accuracy when all three modalities are fused. While advanced fusion techniques like EmbraceNet were tested, they failed to outperform simpler late fusion methods. These findings highlight the effectiveness of straightforward and interpretable multimodal fusion techniques for biometric recognition, providing a balance between performance and model complexity.
|
| Pauza |
| Papers |
D. Kerman (Voltify d.o.o., Čakovec, Croatia), S. Barac (Voltify d.o.o., Zagreb, Croatia), D. Sabolić (Croatian Transmission System Operator, Ltd., Zagreb, Croatia)
Exploring the Operational Dynamics and Structural Properties of Croatia’s EV Charging Network
This study examines key factors influencing the deployment and operation of electric vehicle (EV) charging stations in Croatia, focusing on power capacity, operator types, and geographic distribution. Analysis reveals significant differences in station capacities, with highway locations exhibiting substantially higher power levels, highlighting their importance for long-distance travel. Using complex network theory, statistical analysis, and machine learning, we identify patterns in infrastructure deployment, usage trends, and potential network vulnerabilities. Our findings emphasize the need for strategic planning to optimize EV charging infrastructure, ensuring reliability and capacity for future demand.
|
L. Jakić, I. Botički, A. Sović (Faculty of Electrical Engineering and Computing University of Zagreb, Zagreb, Croatia)
Bridging Virtual and Physical Robotics: An AI-Driven Educational Platform Using NVIDIA Omniverse Isaac Sim and Jetson Orin Nano
This paper presents the development of a platform for educational robotics with artificial intelligence using NVIDIA Omniverse Isaac Sim, a custom robot based on NVIDIA Jetson Orin Nano and ROS2 (Robot Operating System 2) as a standardized development framework. Our research addresses the design and implementation of the system architecture for the platform, with the goal of being simple to use by the students, powerful enough to run machine learning algorithms faster than current state-of-the-art platforms, and ensuring that students’ code is transferable between virtual and physical environments. Previous research within the STEM area highlights multiple examples of using hardware such as Arduino boards or Raspberry Pi for educational robotics lacking the ability to run intensive machine learning algorithms on hardware such as NVIDIA Jetson Orin Nano. The proposed system architecture demonstrates NVIDIA Isaac Sim’s potential in educational robotics by enabling students to experience how their own solutions operate in both virtual and physical learning environments.
|
H. Markulin (RIT Croatia, Zagreb, Croatia), A. Radovan (Algebra University College, Zagreb, Croatia), B. Mihaljević (RIT Croatia, Zagreb, Croatia), K. Bilić (Zagreb University of Applied Sciences, Zagreb, Croatia)
Using ChatGPT in Designing Web Applications Resistant to SQL Injection Attacks
This paper investigates leveraging ChatGPT as a tool for testing web applications resilient to SQL injection attacks. The research begins with a definition of the problem domain, identifying key entities and attributes relevant to the database model. Subsequently, the web application analysis is conducted using different version of ChatGPT, guided by best practices for preventing SQL injection vulnerabilities. The study evaluates the implemented web application based on Spring Framework, analyzing their compliance with security principles and effectiveness in mitigating risks associated with SQL injection. By examining ChatGPT's role in integrating secure coding practices during the database design phase, the research aims to assess its utility as a collaborative and intelligent assistant in the cybersecurity domain. The findings provide insights into the strengths and limitations of AI-driven cybersecurity modeling for enhancing system security against injection-based exploits.
|
M. Drčić, D. Delija (Tehničko veleučilište u Zagrebu, Zagreb, Croatia)
Mrežna akvizicija dokaza u stvarnom vremenu: Usporedba Linux i Windows alata za udaljeno prikupljanje digitalnih dokaza
Digitalna forenzika u stvarnom vremenu jedno je od područja u kojem se alati poput dd i dc3dd u Linuxu te PowerShell naredbe u Windows okruženju koriste za prikupljanje digitalnih dokaza putem mreže. Ovaj rad istražuje kako koristiti dd i dc3dd za izradu disk imagea preko mreže te kako postići slične rezultate korištenjem PowerShell naredbi kao što su Enter-PSSession i udaljeno mapiranje diskova. Uspoređuju se prednosti i izazovi oba pristupa, nudeći praktične primjere za stručnjake koji rade u hibridnim okruženjima. Cilj je omogućiti dublje razumijevanje mrežnih akvizicijskih metoda te istaknuti najbolje prakse za njihovu implementaciju.
|
|
Basic information:
Chairs:
Pavle Ergović (Croatia), Dubravko Sabolić (Croatia), Filip Srdić (Croatia)
Registration / Fees:
|
REGISTRATION / FEES
|
Price in EUR
|
EARLY BIRD
Up to 23 May 2025 |
REGULAR
From 24 May 2025 |
| Members of MIPRO and IEEE |
270 |
297 |
| Students (undergraduate and graduate), primary and secondary school teachers |
150 |
165 |
| Others |
300 |
330 |
The student discount doesn't apply to PhD students.
NOTE FOR AUTHORS: In order to have your paper published, it is required that you pay at least one registration fee for each paper. Authors of 2 or more papers are entitled to a 10% discount.
Contact:
Pavle Ergović
KSET
Faculty of Electrical Engineering and Computing
Unska 3
HR-10000 Zagreb, Croatia
GSM: +385 99 417 3482
E-mail: pavleergovic@gmail.com
The best papers will get a special award.
Accepted papers will be published in the ISSN registered conference proceedings. Papers in English presented at the conference will be submitted for inclusion in the IEEE Xplore Digital Library.
Location:
 Opatija is the leading seaside resort of the Eastern Adriatic and one of the most famous tourist destinations on the Mediterranean. With its aristocratic architecture and style, Opatija has been attracting artists, kings, politicians, scientists, sportsmen, as well as business people, bankers and managers for more than 180 years.
The tourist offer in Opatija includes a vast number of hotels, excellent restaurants, entertainment venues, art festivals, superb modern and classical music concerts, beaches and swimming pools – this city satisfies all wishes and demands.
Opatija, the Queen of the Adriatic, is also one of the most prominent congress cities in the Mediterranean, particularly important for its ICT conventions, one of which is MIPRO, which has been held in Opatija since 1979, and attracts more than a thousand participants from over forty countries. These conventions promote Opatija as one of the most desirable technological, business, educational and scientific centers in South-eastern Europe and the European Union in general.
For more details, please visit www.opatija.hr and visitopatija.com.
|
|
|
|
| Currently there are no news |
|
|
|
|
|