Vitaliy Ivanovich Gorbenko
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Vitaliy Ivanovich Gorbenko has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
A. Gorban
1 shared publicationA. N. Gorban
Classical Private University
1 shared publicationAnastasiia Kryvokhata
Zaporizhzhia National University
1 shared publicationJ. Shvets
1 shared publication- Oleksandr Zhukov
Запорізький національний університет
1 shared publication - Oleksii Kudin
Zaporizhzhia National University
1 shared publication
Latest Publications
- Open Access
- 1 January 2025
- Published by Zaporizhzhia National University in Computer Science and Applied Mathematics
- No. 1,p. 111-118
OPTIMIZING OBJECT DETECTION PERFORMANCE THROUGH IMAGE ENHANCEMENT FOR BEE IMAGES
Neural networks and deep learning, being among the technologies experiencing daily growing interest, have found practical application, particularly in bee monitoring. The use of neural networks for identifying bees in images can provide non-invasive monitoring of the condition of individual bees and, consequently, reduce stress on the colony. However, the accuracy of object detection algorithms can be significantly affected by variations in monitoring conditions, such as weather conditions, lighting, or the use of different cameras. Thus, image enhancement can be an important pre-processing step for optimizing object detection algorithms.This study utilizes various image enhancement and color normalization methods to improve the performance of the SSD MobileNet model in detecting the full bodies of bees in images. The data source consists of photos taken with a smartphone, which were subsequently divided into sections to fit within the model's acceptable identification limits. The study examines the effectiveness of three image enhancement methods: Contrast Limited Adaptive Histogram Equalization (CLAHE), a regular histogram equalization algorithm, and Protobuf pipeline functions. Each approach is evaluated on different image sets, including photos of normal beehive sections taken with a smartphone, photos with randomly altered brightness and contrast, and photos with slight blurring. The findings show that image normalization alone does not improve object detection performance on standard bee photos, whereas the CLAHE algorithm proved most effective in other cases, maintaining good results when image brightness and contrast were not optimized. Additionally, it was found that applying blur did not result in an overall increase in the efficiency of identifying bees in photos. Нейронні мережі та глибоке навчання як технології, інтерес до яких з кожним днем зростає, знайшли практичне застосування, зокрема, у моніторингу бджіл. Використання нейронних мереж для ідентифікації бджіл на зображенні можуть забезпечити неінвазивний моніторинг стану окремих бджіл і, як наслідок, зменшити стрес на колонію. Проте на точність алгоритмів виявлення об’єктів можуть значно вплинути варіації умов моніторингу, як-от погодні умови, освітлення або використання різних камер. Таким чином, покращення зображення може бути важливим етапом попередньої обробки для оптимізації алгоритмів виявлення об’єктів. У цьому дослідженні використовуються різноманітні методи покращення зображення та нормалізації кольору, щоб покращити продуктивність моделі SSD MobileNet у виявленні повних тіл бджіл на зображеннях. Як джерела даних виступають фотографії зроблені на смартфон, які далі були розділені на секції з метою вкладання в допустимі ліміти ідентифікації моделі. У дослідженні розглядається ефективність трьох методів покращення зображення: CLAHE, регулярного алгоритму вирівнювання гістограми та функцій конвеєра Protobuf. Кожен підхід оцінюється на різних наборах зображень, включно з фотографіями звичайних секцій бджолиного вулика, знятих на смартфон, фотографіями з довільно зміненою яскравістю та контрастністю та фотографіями з легким розмиттям. Висновки показують, що нормалізація зображення сама по собі не покращує продуктивність виявлення об’єктів на звичайних фотографіях бджіл, тоді як алгоритм CLAHE виявився найефективнішим в інших випадках, зберігаючи хороші результати у випадках, коли яскравість і контраст зображення не були оптимізовані.Також було виявлено, що застосування розмиття не дало загального приросту в ефективності ідентифікації бджіл на фото.
- 1 January 2025
- Published by Zaporizhzhia National University in Computer Science and Applied Mathematics
- No. 1,p. 111-118
Open AccessOPTIMIZING OBJECT DETECTION PERFORMANCE THROUGH IMAGE ENHANCEMENT FOR BEE IMAGES
Neural networks and deep learning, being among the technologies experiencing daily growing interest, have found practical application, particularly in bee monitoring. The use of neural networks for identifying bees in images can provide non-invasive monitoring of the condition of individual bees and, consequently, reduce stress on the colony. However, the accuracy of object detection algorithms can be significantly affected by variations in monitoring conditions, such as weather conditions, lighting, or the use of different cameras. Thus, image enhancement can be an important pre-processing step for optimizing object detection algorithms.This study utilizes various image enhancement and color normalization methods to improve the performance of the SSD MobileNet model in detecting the full bodies of bees in images. The data source consists of photos taken with a smartphone, which were subsequently divided into sections to fit within the model's acceptable identification limits. The study examines the effectiveness of three image enhancement methods: Contrast Limited Adaptive Histogram Equalization (CLAHE), a regular histogram equalization algorithm, and Protobuf pipeline functions. Each approach is evaluated on different image sets, including photos of normal beehive sections taken with a smartphone, photos with randomly altered brightness and contrast, and photos with slight blurring. The findings show that image normalization alone does not improve object detection performance on standard bee photos, whereas the CLAHE algorithm proved most effective in other cases, maintaining good results when image brightness and contrast were not optimized. Additionally, it was found that applying blur did not result in an overall increase in the efficiency of identifying bees in photos. Нейронні мережі та глибоке навчання як технології, інтерес до яких з кожним днем зростає, знайшли практичне застосування, зокрема, у моніторингу бджіл. Використання нейронних мереж для ідентифікації бджіл на зображенні можуть забезпечити неінвазивний моніторинг стану окремих бджіл і, як наслідок, зменшити стрес на колонію. Проте на точність алгоритмів виявлення об’єктів можуть значно вплинути варіації умов моніторингу, як-от погодні умови, освітлення або використання різних камер. Таким чином, покращення зображення може бути важливим етапом попередньої обробки для оптимізації алгоритмів виявлення об’єктів. У цьому дослідженні використовуються різноманітні методи покращення зображення та нормалізації кольору, щоб покращити продуктивність моделі SSD MobileNet у виявленні повних тіл бджіл на зображеннях. Як джерела даних виступають фотографії зроблені на смартфон, які далі були розділені на секції з метою вкладання в допустимі ліміти ідентифікації моделі. У дослідженні розглядається ефективність трьох методів покращення зображення: CLAHE, регулярного алгоритму вирівнювання гістограми та функцій конвеєра Protobuf. Кожен підхід оцінюється на різних наборах зображень, включно з фотографіями звичайних секцій бджолиного вулика, знятих на смартфон, фотографіями з довільно зміненою яскравістю та контрастністю та фотографіями з легким розмиттям. Висновки показують, що нормалізація зображення сама по собі не покращує продуктивність виявлення об’єктів на звичайних фотографіях бджіл, тоді як алгоритм CLAHE виявився найефективнішим в інших випадках, зберігаючи хороші результати у випадках, коли яскравість і контраст зображення не були оптимізовані.Також було виявлено, що застосування розмиття не дало загального приросту в ефективності ідентифікації бджіл на фото.
- 1 May 2020
- Published by The Electrochemical Society in ECS Meeting Abstracts
- Vol. MA2020-01 (26) , 1853
Developing a Deep Learning Sound Classification System for a Smart Farming
The Internet of Things (IoT) and Machine Learning (ML) are a promising technologies for automation in the different domains e.g. Smart Farming. Such systems could be used for determining noise emission, animal detection and classification, monitoring states of bees in beehive and etc. As rule the characteristics of that systems are some controversial: compactness (deploying on the single-board computers) and real time classification and prediction. Therefore, the software for IoT and ML applications should be sufficiently efficient and not demanding on large computing resources. This work are focused on machine hearing system allows to classify natural sounds in the Smart Farming software application. The original acoustic signal is converted to the frames of a certain length. Receiving a compact representation of the acoustic characteristics of a signal is the aim of the feature extraction stage. This stage exploits special coefficients such as Zero-crossing rate, Spectrum shape, Short-Time Fourier Transform and Mel-frequency cepstral coefficients. Audio classification traditionally involves such machine learning methods as K-means, support vector machine (SVM), decision trees etc. Deep neural networks can be used on both raw acoustic signal and features extracted from it. The accuracy estimation stage deploys quality assessment methods. The data preprocessing stage includes calculating of mel frequency cepstral coefficient (MFCC) for the giving sound files. This approach allows to unify and simplify the sound files presentation in the memory. Further MFCC arrays are feeding to the convolutional neural net. It is important at this stage to configure the network optimally for the most compact storage. This is due to the need to use platforms such as Raspberry Pi to deploying neural network. For neural network software implementation Python library Keras is used. For data preprocessing Python library Librosa is used. The hyperparameters of the neural network have been defined in computational experiments and the optimal combination is two Conv2D layers and three Dense layers. The accuracy of model predictions for each class has been examined by using a confusion matrix and the satisfactory classification accuracy has been defined. It has been established the convolutional neural network ensemble built to solve the problem of acoustic data classification is quite effective. The accuracy of the model on the test data is 95%. It should be noted that the neural network structure with only two packets of convolution-activation-sub-sampling layers was sufficient to solve this problem. The system can be used in Smart Farming applications to filter out unnecessary sounds.
- 1 May 2020
- Published by The Electrochemical Society in ECS Meeting Abstracts
- Vol. MA2020-01 (26) , 1853
Developing a Deep Learning Sound Classification System for a Smart Farming
The Internet of Things (IoT) and Machine Learning (ML) are a promising technologies for automation in the different domains e.g. Smart Farming. Such systems could be used for determining noise emission, animal detection and classification, monitoring states of bees in beehive and etc. As rule the characteristics of that systems are some controversial: compactness (deploying on the single-board computers) and real time classification and prediction. Therefore, the software for IoT and ML applications should be sufficiently efficient and not demanding on large computing resources. This work are focused on machine hearing system allows to classify natural sounds in the Smart Farming software application. The original acoustic signal is converted to the frames of a certain length. Receiving a compact representation of the acoustic characteristics of a signal is the aim of the feature extraction stage. This stage exploits special coefficients such as Zero-crossing rate, Spectrum shape, Short-Time Fourier Transform and Mel-frequency cepstral coefficients. Audio classification traditionally involves such machine learning methods as K-means, support vector machine (SVM), decision trees etc. Deep neural networks can be used on both raw acoustic signal and features extracted from it. The accuracy estimation stage deploys quality assessment methods. The data preprocessing stage includes calculating of mel frequency cepstral coefficient (MFCC) for the giving sound files. This approach allows to unify and simplify the sound files presentation in the memory. Further MFCC arrays are feeding to the convolutional neural net. It is important at this stage to configure the network optimally for the most compact storage. This is due to the need to use platforms such as Raspberry Pi to deploying neural network. For neural network software implementation Python library Keras is used. For data preprocessing Python library Librosa is used. The hyperparameters of the neural network have been defined in computational experiments and the optimal combination is two Conv2D layers and three Dense layers. The accuracy of model predictions for each class has been examined by using a confusion matrix and the satisfactory classification accuracy has been defined. It has been established the convolutional neural network ensemble built to solve the problem of acoustic data classification is quite effective. The accuracy of the model on the test data is 95%. It should be noted that the neural network structure with only two packets of convolution-activation-sub-sampling layers was sufficient to solve this problem. The system can be used in Smart Farming applications to filter out unnecessary sounds.
- 1 May 2020
- Published by The Electrochemical Society in ECS Meeting Abstracts
- Vol. MA2020-01 (23) , 1361
Study of Graphene-like Indium Phosphide
The semiconductor compounds in a state of graphene-like structure are the interesting object for both fundamental studies and technology. The two-dimensional material have unique electronic, structural, chemical and physical properties that plays a significant role for technology of new devices. Therefore the obtaining of various properties of that graphene-like materials and the understanding of the nature of chemical interaction with various species are important. In this work for the geometry optimisation graphene-like structure of indium phosphide the GAMESS program package [1] have been used. The electronic properties of g-InP has been obtained by ABINIT program package [2]. Such species as hydrogen, oxygen and carbon are a part of many natural compositions of the environment. In this work the preferred adsorption sites for the H, O and C adatoms have been defined through the potential-energy profiles of its interacting with g-InP sheet. All ab initio calculations reported in this work were carried out using the Hartree-Fock method of GAMESS program package. The GAMESS was run on computers cluster of V.M.Glushkov Institute of Cybernetics of NAS of Ukraine [3]. Both STO-3G and 3-21G Pople style basis sets were used for C, H and O atoms. The g-InP sheet has been presented by cluster model with 48 In-atoms and 48 P-atoms in the hexagonal cells and 24 hydrogen atoms on the cluster perimeter. In the atomic-force probe technique the atom-probe (H, O or C) is moving above the surface of sheet on fixed distance without any optimization of its geometry. As a result the potential-energy (PE) profiles of interaction of the various atoms-probe with g-InP sheet have been obtained. [1] Gordon M.S., Schmidt M.W. // in "Theory and Applications of Computational Chemistry: the first forty years" by ed. C.E.Dykstra. - Elsevier, Amsterdam, 2005, p.1167-1189. [2] https://docs.abinit.org/topics/features/ [3] http://www.icyb.kiev.ua/s/189/ua/klaster_ik.html Figure 1
- 1 May 2020
- Published by The Electrochemical Society in ECS Meeting Abstracts
- Vol. MA2020-01 (23) , 1361
Study of Graphene-like Indium Phosphide
The semiconductor compounds in a state of graphene-like structure are the interesting object for both fundamental studies and technology. The two-dimensional material have unique electronic, structural, chemical and physical properties that plays a significant role for technology of new devices. Therefore the obtaining of various properties of that graphene-like materials and the understanding of the nature of chemical interaction with various species are important. In this work for the geometry optimisation graphene-like structure of indium phosphide the GAMESS program package [1] have been used. The electronic properties of g-InP has been obtained by ABINIT program package [2]. Such species as hydrogen, oxygen and carbon are a part of many natural compositions of the environment. In this work the preferred adsorption sites for the H, O and C adatoms have been defined through the potential-energy profiles of its interacting with g-InP sheet. All ab initio calculations reported in this work were carried out using the Hartree-Fock method of GAMESS program package. The GAMESS was run on computers cluster of V.M.Glushkov Institute of Cybernetics of NAS of Ukraine [3]. Both STO-3G and 3-21G Pople style basis sets were used for C, H and O atoms. The g-InP sheet has been presented by cluster model with 48 In-atoms and 48 P-atoms in the hexagonal cells and 24 hydrogen atoms on the cluster perimeter. In the atomic-force probe technique the atom-probe (H, O or C) is moving above the surface of sheet on fixed distance without any optimization of its geometry. As a result the potential-energy (PE) profiles of interaction of the various atoms-probe with g-InP sheet have been obtained. [1] Gordon M.S., Schmidt M.W. // in "Theory and Applications of Computational Chemistry: the first forty years" by ed. C.E.Dykstra. - Elsevier, Amsterdam, 2005, p.1167-1189. [2] https://docs.abinit.org/topics/features/ [3] http://www.icyb.kiev.ua/s/189/ua/klaster_ik.html Figure 1
- 2 April 2013
- Published by The Electrochemical Society in ECS Transactions
- Vol. 45 (20) , 185-190
Study of H, O and C Adatoms Surface Diffusion on Graphene by Computer Simulation
The ab initio methods of GAMESS package have been used to calculate the adsorption and the surface diffusion properties of H, O and C adatoms on graphene sheet. The graphene sheet was presented by cluster model with hydrogen atoms on perimeter. The potential-energy profiles of interaction of the adatoms with graphene have been obtained. The essential difference of the profiles and the energy barriers has been found for all adsorbate species.
- 2 April 2013
- Published by The Electrochemical Society in ECS Transactions
- Vol. 45 (20) , 185-190
Study of H, O and C Adatoms Surface Diffusion on Graphene by Computer Simulation
The ab initio methods of GAMESS package have been used to calculate the adsorption and the surface diffusion properties of H, O and C adatoms on graphene sheet. The graphene sheet was presented by cluster model with hydrogen atoms on perimeter. The potential-energy profiles of interaction of the adatoms with graphene have been obtained. The essential difference of the profiles and the energy barriers has been found for all adsorbate species.
- Open Access
- 25 July 2012
- Published by National University "Zaporizhzhia Polytechnic" in Radio Electronics, Computer Science, Control
- No. 1
THERMAL DECOMPOSITION OF INDIUM PHOSPHIDE IN VACUUM AND ATOMIC HYDROGEN ENVIRONMENT
The thermal decomposition of indium phosphide has been investigated by Auger-electron spectroscopy and mass-spectroscopy. Scanning electron microscopy has been used for study of indium islands growth on surface of the compound semiconductor. The role of atomic hydrogen in processes of decomposition and growth of metallic islands was determined by comparing with these processes under vacuum.
- 25 July 2012
- Published by National University "Zaporizhzhia Polytechnic" in Radio Electronics, Computer Science, Control
- No. 1
Open AccessTHERMAL DECOMPOSITION OF INDIUM PHOSPHIDE IN VACUUM AND ATOMIC HYDROGEN ENVIRONMENT
The thermal decomposition of indium phosphide has been investigated by Auger-electron spectroscopy and mass-spectroscopy. Scanning electron microscopy has been used for study of indium islands growth on surface of the compound semiconductor. The role of atomic hydrogen in processes of decomposition and growth of metallic islands was determined by comparing with these processes under vacuum.