CSIT - 2024 рік
Постійне посилання зібрання
Переглянути
Перегляд CSIT - 2024 рік за Ключові слова "004.9"
Зараз показуємо 1 - 7 з 7
Результатів на сторінці
Налаштування сортування
Документ Cyber-physical system for monitoring the environment for allergens using geolocation data(Хмельницький національний університет, 2024) Hovorushchenko, T.; Voevudskyi, Y.; Ivanov, O.; Voichur, O.Документ Method for improving the performance of convolutional neural networks using an accelerator(Хмельницький національний університет, 2024) Isaiev, T.; Kysil , T.The effectiveness of convolutional neural networks (CNNs) has been demonstrated across various fields, including computer vision, natural language processing, medical imaging, and autonomous systems. However, achieving high performance in CNNs is not only a matter of model design but also of optimizing the training and inference processes. Using accelerators like the Google Coral TPU provides significant improvements in both computational efficiency and overall model performance. This paper focuses on the integration of the Coral TPU to enhance CNN performance by speeding up computations, reducing latency, and enabling real-time deployment. Training deep learning models, particularly CNNs, is computationally intensive. Traditional CPUs or GPUs can take hours or even days to train large networks on complex data. The accelerator offloads these intensive tasks, allowing the host machine to focus on other operations and making training more efficient. This enables researchers to experiment with multiple architectures and hyperparameters within shorter cycles, thereby improving the model's accuracy and robustness. CNNs are widely deployed in edge computing scenarios where real-time predictions are critical, such as in robotics, autonomous vehicles, and smart surveillance systems.Unlike traditional cloud-based solutions, where models are executed remotely and suffer from network delays, the Coral TPU ensures low-latency predictions directly on the device, making it ideal for timesensitive applications. Another key advantage of using accelerators like Coral TPU is the ability to efficiently handle optimized and lightweight models. These optimized models are well-suited for the Coral TPU’s architecture, allowing developers to deploy high-performing networks even on resource-constrained devices. The TPU’s ability to handle quantized models with minimal loss in accuracy further enhances the CNN’s practical usability across various domains. The Coral TPU is designed to minimize power consumption, making it an ideal solution for battery-powered or energyconstrained devices. This energy efficiency ensures that CNNs can run continuously on devices like drones, IoT sensors, or mobile platforms without exhausting their power supply. Additionally, the scalability of the TPU makes it easy to deploy multiple accelerators in parallel, further improving throughput for applications that require processing high volumes of data, such as realtime video analysis. The Coral TPU also facilitates on-device learning, where models can be incrementally updated based on new data without requiring a full retraining session. This feature is particularly useful in dynamic environments, such as autonomous vehicles or security systems, where the model needs to adapt quickly to new conditions. With the TPU handling the computational workload, CNNs can be fine-tuned on the device, ensuring they remain accurate and responsive over time.Документ Method for interpreting decisions made by deep learning models(Хмельницький національний університет, 2024) Slobodzian, V.; Barmak, O.The use of artificial intelligence (AI) in medical diagnostics opens new opportunities for analyzing complex medical images and optimizing diagnostic processes. One of the key challenges remains the interpretation of results obtained through AI systems, particularly in medical practice, where ensuring transparency and clarity of decision-making is critically important. This study proposes a method for visualizing and interpreting the results of cardiac disease classification based on MRI image analysis using deep learning models. The primary goal of the research is to explain AI-driven decisions in a convenient and understandable format for physicians, contributing to the reduction of subjectivity in clinical practice. During the research, approaches were developed for visualizing key groups of medical indicators, such as heart volumes, ejection fraction, myocardial wall thickness, and volume-to-mass ratios. The study describes numerical metrics commonly used in medical practice. Fifteen key medical metrics were identified and grouped into corresponding categories for effective representation of essential medical indicators. Various visualization forms were utilized to ensure intuitive data presentation: pie charts to demonstrate ratios, the 17-segment myocardial model for analyzing wall thickness, and numerical indicators for accurately displaying volumes and ejection fraction. This approach allows physicians to quickly assess structural changes in the heart and make informed conclusions. The proposed method aims to enhance transparency and trust in AI by providing comprehensible data representation, reducing the risks of subjective interpretation and cognitive biases. The results indicate that using such visualizations can significantly facilitate clinical decision-making, improve diagnostic accuracy, and standardize approaches to medical data analysis.Документ Method of creating custom dataset to train convolutional neural network(Хмельницький національний університет, 2024) Isaiev, T.; Kysil , T.The task of creating and developing custom datasets for training convolutional neural networks (CNNs) is essential due to the increasing adoption of deep learning across industries. CNNs have become fundamental tools for various applications, including computer vision, natural language processing, medical imaging, and autonomous systems. However, the success of a CNN depends heavily on the quality and relevance of the data it is trained on. The datasets used to train these models must be diverse, representative of the task at hand, and of sufficient quality to capture the underlying patterns that the CNN needs to learn. Thus, building custom datasets that align with the specific objectives of a neural network plays a critical role in enhancing the performance and generalization capability of the trained model. This paper focuses on developing a method and subsystem for generating high-quality custom datasets tailored to CNNs. The aim is to provide a framework that automates and streamlines the processes involved in data collection, preprocessing, augmentation, annotation, and validation. Moreover, the method integrates tools that allow the dataset to evolve over time, incorporating new data to adapt to changing requirements or environments, making the system flexible and scalable. The process of creating a dataset begins with the acquisition of raw data. The data can come from various sources such as images from cameras, videos, sensor feeds, open data repositories, or proprietary datasets. A key consideration during data collection is ensuring that the samples cover the full range of conditions or classes the CNN will encounter in production. For example, in an object recognition task, it is essential to collect images from diverse environments, lighting conditions, and angles to train the model effectively. Ensuring variability in the dataset increases the model's ability to generalize, reducing the risk of poor performance on unseen data. Data augmentation is a critical step in building a robust dataset, particularly when the size of the dataset is limited. Augmentation techniques introduce variability into the dataset by artificially modifying the existing samples, thereby simulating a wider range of conditions. This helps the CNN generalize better and prevents overfitting. In essence, it allows the model to experience different perspectives and distortions of the same data, strengthening its adaptability to real-world scenarios. Annotation involves labeling the data samples with the correct class or category information. Depending on the task, annotations may include bounding boxes for object detection, segmentation masks for semantic segmentation, or class labels for classification tasks. The importance of well-annotated data cannot be overstated, as CNNs rely on this labeled information to understand the relationships between input data and the desired output predictions.Документ Overview of the methods and tools for environmental components monitoring(Хмельницький національний університет, 2024) Hovorushchenko, T.; Bachuk, V.; Hnatchuk, Y.; Zasornova, I.; Bouhissi, H. E.Monitoring of environmental components is an important process for determining the level of pollution and tracking changes in the environment, and plays a key role in ensuring the health and comfort of residents, as well as in preserving the environment. Continuous monitoring of environmental components is key to ensuring human health, protecting nature and reducing the negative impact on the climate and ecosystems, as well as achieving sustainable development. In order to combat environmental pollution, it is necessary to implement effective measures to limit emissions of harmful substances, use environmentally friendly technologies and green solutions in all sectors of the economy, and raise public awareness of the problem of environmental pollution. From the analysis of the sources reviewed, a pattern was identified that the information technologies mainly used to monitor environmental components are either Internet of Things (IoT) technologies using modern sensors and data transmission components or artificial intelligence technologies such as computer vision. Less commonly, the use of robots, UAVs, and digital twins is being traced. Based on a critical analysis of methods and tools for environmental components monitoring, there is a need to develop such methods and tools for environmental components monitoring that would: perform cheaper and more versatile environmental components monitoring than existing analogues, but at the same time have no less accuracy and speed; monitor the state of the environment; identify sources of environmental pollution; warn of environmental disasters; assess the state of natural resources; support environmental decision-making; collect and analyze various environmental indicators in real time; assessed the level of quality and safety of environmental components, which will allow immediate response to quality changes and promptly take the necessary measures, etc., which will be the focus of the authors' further efforts.Документ Overview of the methods and tools for environmental components monitoring(Хмельницький національний університет, 2024) Hovorushchenko, T.; Bachuk, V.; Hnatchuk, Y.; Zasornova, I.; El, Bouhissi H.Monitoring of environmental components is an important process for determining the level of pollution and tracking changes in the environment, and plays a key role in ensuring the health and comfort of residents, as well as in preserving the environment. Continuous monitoring of environmental components is key to ensuring human health, protecting nature and reducing the negative impact on the climate and ecosystems, as well as achieving sustainable development. In order to combat environmental pollution, it is necessary to implement effective measures to limit emissions of harmful substances, use environmentally friendly technologies and green solutions in all sectors of the economy, and raise public awareness of the problem of environmental pollution. From the analysis of the sources reviewed, a pattern was identified that the information technologies mainly used to monitor environmental components are either Internet of Things (IoT) technologies using modern sensors and data transmission components or artificial intelligence technologies such as computer vision. Less commonly, the use of robots, UAVs, and digital twins is being traced. Based on a critical analysis of methods and tools for environmental components monitoring, there is a need to develop such methods and tools for environmental components monitoring that would: perform cheaper and more versatile environmental components monitoring than existing analogues, but at the same time have no less accuracy and speed; monitor the state of the environment; identify sources of environmental pollution; warn of environmental disasters; assess the state of natural resources; support environmental decision-making; collect and analyze various environmental indicators in real time; assessed the level of quality and safety of environmental components, which will allow immediate response to quality changes and promptly take the necessary measures, etc., which will be the focus of the authors' further efforts.Документ Subsystem for monitoring atmospheric air quality in the cyberphysical system "Smart City"(Хмельницький національний університет, 2024) Hovorushchenko, T.; Baranovskyi, V.; Hnatchuk, A.; Ivanov, O.The task of designing and developing a cyber-physical system "Smart City" is currently relevant for Ukraine. This study is devoted to the development of a method and subsystem for monitoring atmospheric air quality in the cyber-physical system "Smart City". The article develops a method for monitoring atmospheric air quality, which forms the basis for effective monitoring of atmospheric air quality in the cyber-physical system "Smart City" and allows making informed decisions on warning residents about the danger with recommendations for protecting their health. The developed subsystem for monitoring atmospheric air quality in the cyber-physical system “Smart City” collects data from the installed sensors of air humidity, air temperature, dust content in the air, including particles PM2.5, PM10, air radiation background, air pollution level by nitrogen oxides, air pollution level by sulfur, air pollution level by carbon compounds, air pollution level by greenhouse gases CO, CO2, NH3, NO, real-time transmission of the collected data to the data processing server, real-time processing and analysis of the received data using various analytical methods, visualization of the air quality monitoring results in the form of a city map with n districts displaying all air parameters. The user can select the air parameters of interest in the mobile application of the cyber-physical system. After selecting such parameters, the visualization of the air quality monitoring results is adapted to the user's needs: the measured value of the parameter selected by the user is displayed on the image of the district on the city map, and the mobile application displays a sound signal in the background and a flashing sign on the image of the district on the city map in the application, which signals a danger in this area of the city; clicking on this sign displays a notification on the screen about the indicator for which there is a danger and recommendations for protecting the health of residents in this case.