Yunus Egi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Yunus Egi has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
Engin Eyceyurt
Electrical and Electronics Engineering Department, Nevsehir Haci Bektas Veli University, Nevsehir, Turkey
3 shared publicationsCarlos E. Otero
Electrical and Computer Engineering, Florida Institute of Technology, Melbourne, USA
2 shared publications- Engin Eyceyurt
Nevsehir H.B.V. University
2 shared publications - Mortaza Hajyzadeh
Department of Field Crops, Faculty of Agriculture, Sirnak University, Sirnak 73000, Türkiye
2 shared publications - Josko Zec
Florida Institute of Technology
1 shared publication - Mehmet Beşkirli
Department of Medical Education and Informatics, Faculty of Medicine, Karamanoğlu Mehmetbey University, Karaman 70200, Türkiye
1 shared publication
Latest Publications
- Open Access
Real-Time Callus Instance Segmentation in Plant Tissue Culture Using Successive Generations of YOLO Architectures
Callus induction is a complex procedure in plant organ, cell, and tissue culture that underpins processes such as metabolite production, regeneration, and genetic transformation. It is important to monitor callus formation alongside subjective evaluations, which require labor-intensive care. In this research, the first curated lentil (Lens culinaris) callus dataset for instance segmentation was experimentally generated using three genotypes as one data set: Firat-87, Cagil, and Tigris. Leaf explants were cultured on MS medium fortified with different concentrations of gross regulators of BA and NAA to induce callus formation. Three biologically relevant stages, the leaf stage, the green callus, and the necrosis callus, were produced. During this process, 122 high-resolution images were obtained, resulting in 1185 total annotations across them. The dataset was evaluated across four successive generations (v5/7/8/11) of YOLO deep learning models under identical conditions using , coefficient, , , and , together with efficiency metrics including parameter counts, FLOPs, and inference speed. The results show that anchor-based variants (YOLOv5/7) relied on predefined priors and showed limited boundary precision, whereas anchor-free designs (YOLOv8/11) used decoupled heads and direct center/boundary regression that provided clear advantages for callus structures. YOLOv8 reached the highest instance segmentation precision with mAP50@0.855, while it matched the accuracy with greater efficiency and achieved real-time inference with 166 FPS.Open AccessReal-Time Callus Instance Segmentation in Plant Tissue Culture Using Successive Generations of YOLO Architectures
Callus induction is a complex procedure in plant organ, cell, and tissue culture that underpins processes such as metabolite production, regeneration, and genetic transformation. It is important to monitor callus formation alongside subjective evaluations, which require labor-intensive care. In this research, the first curated lentil (Lens culinaris) callus dataset for instance segmentation was experimentally generated using three genotypes as one data set: Firat-87, Cagil, and Tigris. Leaf explants were cultured on MS medium fortified with different concentrations of gross regulators of BA and NAA to induce callus formation. Three biologically relevant stages, the leaf stage, the green callus, and the necrosis callus, were produced. During this process, 122 high-resolution images were obtained, resulting in 1185 total annotations across them. The dataset was evaluated across four successive generations (v5/7/8/11) of YOLO deep learning models under identical conditions using , coefficient, , , and , together with efficiency metrics including parameter counts, FLOPs, and inference speed. The results show that anchor-based variants (YOLOv5/7) relied on predefined priors and showed limited boundary precision, whereas anchor-free designs (YOLOv8/11) used decoupled heads and direct center/boundary regression that provided clear advantages for callus structures. YOLOv8 reached the highest instance segmentation precision with mAP50@0.855, while it matched the accuracy with greater efficiency and achieved real-time inference with 166 FPS. - Open Access
- 1 March 2023
- Published by Igdir University in Iğdır Üniversitesi Fen Bilimleri Enstitüsü Dergisi
- Vol. 13 (1) , 22-32
YOLO V7 and Computer Vision-Based Mask-Wearing Warning System for Congested Public Areas
The impact of Covid 19 cases is increasing worldwide due to not complying with social distancing and mask-wearing rules in congested areas such as hospitals, schools, and malls where people have to be together. Although the authorities have taken various precautions to prevent not wearing masks, it is challenging to inspect masks in crowded areas. People who do not wear masks can be unnoticed by visual inspections, which is a critical factor in the increase of the epidemic. This study aims to create an Artificial Intelligence (AI) based mask inspection system with the YOLO V7 deep learning method to ensure that overcrowded public areas are protected from the Covid-19 epidemic.
- 1 March 2023
- Published by Igdir University in Iğdır Üniversitesi Fen Bilimleri Enstitüsü Dergisi
- Vol. 13 (1) , 22-32
Open AccessYOLO V7 and Computer Vision-Based Mask-Wearing Warning System for Congested Public Areas
The impact of Covid 19 cases is increasing worldwide due to not complying with social distancing and mask-wearing rules in congested areas such as hospitals, schools, and malls where people have to be together. Although the authorities have taken various precautions to prevent not wearing masks, it is challenging to inspect masks in crowded areas. People who do not wear masks can be unnoticed by visual inspections, which is a critical factor in the increase of the epidemic. This study aims to create an Artificial Intelligence (AI) based mask inspection system with the YOLO V7 deep learning method to ensure that overcrowded public areas are protected from the Covid-19 epidemic.
- 13 December 2022
- Published by Springer Nature in Neural Computing and Applications
- Vol. 35 (10) , 7611-7633
An efficient hybrid-based charged system search algorithm for active filter design
- 13 December 2022
- Published by Springer Nature in Neural Computing and Applications
- Vol. 35 (10) , 7611-7633
An efficient hybrid-based charged system search algorithm for active filter design
- Open Access
- 23 October 2022
- Published by Springer Nature in EURASIP Journal on Wireless Communications and Networking
- Vol. 2022 (1) , 1-17
Classified 3D mapping and deep learning-aided signal power estimation architecture for the deployment of wireless communication systems
- 23 October 2022
- Published by Springer Nature in EURASIP Journal on Wireless Communications and Networking
- Vol. 2022 (1) , 1-17
Open AccessClassified 3D mapping and deep learning-aided signal power estimation architecture for the deployment of wireless communication systems
- Open Access
- 23 August 2022
- Published by MDPI AG in Agriculture
- Vol. 12 (9) , 1290
Drone-Computer Communication Based Tomato Generative Organ Counting Model Using YOLO V5 and Deep-Sort
The growth and development of generative organs of the tomato plant are essential for yield estimation and higher productivity. Since the time-consuming manual counting methods are inaccurate and costly in a challenging environment, including leaf and branch obstruction and duplicate tomato counts, a fast and automated method is required. This research introduces a computer vision and AI-based drone system to detect and count tomato flowers and fruits, which is a crucial step for developing automated harvesting, which improves time efficiency for farmers and decreases the required workforce. The proposed method utilizes the drone footage of greenhouse tomatoes data set containing three classes (red tomato, green tomato, and flower) to train and test the counting model through YOLO V5 and Deep Sort cutting-edge deep learning algorithms. The best model for all classes is obtained at epoch 96 with an accuracy of 0.618 at mAP 0.5. Precision and recall values are determined as 1 and 0.85 at 0.923 and 0 confidence levels, respectively. The F1 scores of red tomato, green tomato, and flower classes are determined as 0.74, 0.56, and 0.61, respectively. The average F1 score for all classes is also obtained as 0.63. Through obtained detection and counting model, the tomato fruits and flowers are counted systematically from the greenhouse environment. The manual and AI-Drone counting results show that red tomato, green tomato, and flowers have 85%, 99%, and 50% accuracy, respectively.
- 23 August 2022
- Published by MDPI AG in Agriculture
- Vol. 12 (9) , 1290
Open AccessDrone-Computer Communication Based Tomato Generative Organ Counting Model Using YOLO V5 and Deep-Sort
The growth and development of generative organs of the tomato plant are essential for yield estimation and higher productivity. Since the time-consuming manual counting methods are inaccurate and costly in a challenging environment, including leaf and branch obstruction and duplicate tomato counts, a fast and automated method is required. This research introduces a computer vision and AI-based drone system to detect and count tomato flowers and fruits, which is a crucial step for developing automated harvesting, which improves time efficiency for farmers and decreases the required workforce. The proposed method utilizes the drone footage of greenhouse tomatoes data set containing three classes (red tomato, green tomato, and flower) to train and test the counting model through YOLO V5 and Deep Sort cutting-edge deep learning algorithms. The best model for all classes is obtained at epoch 96 with an accuracy of 0.618 at mAP 0.5. Precision and recall values are determined as 1 and 0.85 at 0.923 and 0 confidence levels, respectively. The F1 scores of red tomato, green tomato, and flower classes are determined as 0.74, 0.56, and 0.61, respectively. The average F1 score for all classes is also obtained as 0.63. Through obtained detection and counting model, the tomato fruits and flowers are counted systematically from the greenhouse environment. The manual and AI-Drone counting results show that red tomato, green tomato, and flowers have 85%, 99%, and 50% accuracy, respectively.
- Open Access
- 13 April 2022
- Published by MDPI AG in Electronics
- Vol. 11 (8) , 1227
Machine-Learning-Based Uplink Throughput Prediction from Physical Layer Measurements
The uplink (UL) throughput prediction is indispensable for a sustainable and reliable cellular network due to the enormous amounts of mobile data used by interconnecting devices, cloud services, and social media. Therefore, network service providers implement highly complex mobile network systems with a large number of parameters and feature add-ons. In addition to the increased complexity, old-fashioned methods have become insufficient for network management, requiring an autonomous calibration to minimize utilization of the system parameter and the processing time. Many machine learning algorithms utilize the Long-Term Evolution (LTE) parameters for channel throughput prediction, mainly in favor of downlink (DL). However, these algorithms have not achieved the desired results because UL traffic prediction has become more critical due to the channel asymmetry in favor of DL throughput closing rapidly. The environment (urban, suburban, rural areas) affect should also be taken into account to improve the accuracy of the machine learning algorithm. Thus, in this research, we propose a machine learning-based UL data rate prediction solution by comparing several machine learning algorithms for three locations (Houston, Texas, Melbourne, Florida, and Batman, Turkey) and determine the best accuracy among all. We first performed an extensive LTE data collection in proposed locations and determined the LTE lower layer parameters correlated with UL throughput. The selected LTE parameters, which are highly correlated with UL throughput (RSRP, RSRQ, and SNR), are trained in five different learning algorithms for estimating UL data rates. The results show that decision tree and k-nearest neighbor algorithms outperform the other algorithms at throughput estimation. The prediction accuracy with the R2 determination coefficient of 92%, 85%, and 69% is obtained from Melbourne, Florida, Batman, Turkey, and Houston, Texas, respectively.
- 13 April 2022
- Published by MDPI AG in Electronics
- Vol. 11 (8) , 1227
Open AccessMachine-Learning-Based Uplink Throughput Prediction from Physical Layer Measurements
The uplink (UL) throughput prediction is indispensable for a sustainable and reliable cellular network due to the enormous amounts of mobile data used by interconnecting devices, cloud services, and social media. Therefore, network service providers implement highly complex mobile network systems with a large number of parameters and feature add-ons. In addition to the increased complexity, old-fashioned methods have become insufficient for network management, requiring an autonomous calibration to minimize utilization of the system parameter and the processing time. Many machine learning algorithms utilize the Long-Term Evolution (LTE) parameters for channel throughput prediction, mainly in favor of downlink (DL). However, these algorithms have not achieved the desired results because UL traffic prediction has become more critical due to the channel asymmetry in favor of DL throughput closing rapidly. The environment (urban, suburban, rural areas) affect should also be taken into account to improve the accuracy of the machine learning algorithm. Thus, in this research, we propose a machine learning-based UL data rate prediction solution by comparing several machine learning algorithms for three locations (Houston, Texas, Melbourne, Florida, and Batman, Turkey) and determine the best accuracy among all. We first performed an extensive LTE data collection in proposed locations and determined the LTE lower layer parameters correlated with UL throughput. The selected LTE parameters, which are highly correlated with UL throughput (RSRP, RSRQ, and SNR), are trained in five different learning algorithms for estimating UL data rates. The results show that decision tree and k-nearest neighbor algorithms outperform the other algorithms at throughput estimation. The prediction accuracy with the R2 determination coefficient of 92%, 85%, and 69% is obtained from Melbourne, Florida, Batman, Turkey, and Houston, Texas, respectively.
- Open Access
- 15 February 2022
- Published by Springer Nature in Research Square
Classified 3D Mapping and Deep Learning Aided Signal Power Estimation Architecture for the Deployment of Wireless Communication Systems
The traditional wireless communication systems deployment models require expensive and time-consuming procedures, including environment selection(rural, urban, and suburban), drive test data collection, and analysis of the raw data. These procedures mainly utilize stochastic and deterministic approaches for signal strength prediction to locate the optimum cellular tower(eNodeB) position for 4G and 5G systems. Since environment selection is limited by urban, suburban, and rural areas, they do not cover complex macro and micro variations, especially buildings and tree canopies having a higher impact on signal fading due to scattering and absorption. Therefore, they usually end up with high prediction errors. This article proposes an efficient architecture for the deployment of communication systems. The proposed method determines the effect of the environment via extracting tree and building properties by using a classified 3D map and You Only Look Once (YOLO) V5, which is one of the most efficient deep learning algorithms. According to the results, the Mean Average Precision(mAP) 0.5% and mAP 0.95% accuracies are obtained as 0.96 and 0.45, and Image Color Classification(ICC) findings indicate 77.6% accuracy on vegetation detection, especially for tree canopies. Thus, the obtained results significantly improved signal strength prediction with a 3.96% Mean Absolute Percentage Error(MAPE) rate, while other empirical models’ prediction errors fall in the range of 6.07%-15.26%.
- 15 February 2022
- Published by Springer Nature in Research Square
Open AccessClassified 3D Mapping and Deep Learning Aided Signal Power Estimation Architecture for the Deployment of Wireless Communication Systems
The traditional wireless communication systems deployment models require expensive and time-consuming procedures, including environment selection(rural, urban, and suburban), drive test data collection, and analysis of the raw data. These procedures mainly utilize stochastic and deterministic approaches for signal strength prediction to locate the optimum cellular tower(eNodeB) position for 4G and 5G systems. Since environment selection is limited by urban, suburban, and rural areas, they do not cover complex macro and micro variations, especially buildings and tree canopies having a higher impact on signal fading due to scattering and absorption. Therefore, they usually end up with high prediction errors. This article proposes an efficient architecture for the deployment of communication systems. The proposed method determines the effect of the environment via extracting tree and building properties by using a classified 3D map and You Only Look Once (YOLO) V5, which is one of the most efficient deep learning algorithms. According to the results, the Mean Average Precision(mAP) 0.5% and mAP 0.95% accuracies are obtained as 0.96 and 0.45, and Image Color Classification(ICC) findings indicate 77.6% accuracy on vegetation detection, especially for tree canopies. Thus, the obtained results significantly improved signal strength prediction with a 3.96% Mean Absolute Percentage Error(MAPE) rate, while other empirical models’ prediction errors fall in the range of 6.07%-15.26%.
- Open Access
- 2 February 2022
- Published by IntechOpen
3D Point Cloud-Based Tree Canopy Visualization for a Smart Deployment of Mobile Communication Systems
Mobile communication is one of the most important parameters of smart cities in terms of maintaining connectivity and interaction between humans and smart systems. However, In the deployment process of Mobile Communication Systems (MCS), Radio Frequency (RF) engineers use location depended empirical Signal Strength Path Loss (SSPL) models ending up with poor signal strength and slow data connection. This is due to the fact that empirical propagation models usually are restrained by the environment and do not implement state of the art technologies, including Unmanned Aerial Vehicles (UAV), Light Detection and Ranging (LiDAR), Image Processing, and Machine Learning to increase efficiency. Terrains involving buildings, hills, trees, mountains, and human-made structures are considered irregular terrains by telecommunication engineers. Irregular terrains, specifically trees, significantly affect MCS’s efficiency because of their complex pattern resulting in erroneous signal fading via multi-path reflection and absorption. Therefore, a virtual 3D environment is required to extract the required 3D terrain pattern and elevation data from the environment. Once this data is processed in the machine learning algorithm, an adaptive propagation model can be formed and can significantly improve SSPL prediction accuracy for MCS. This chapter presents 3D point cloud visualization via sensor fusion and 2D image color classification techniques, which lead to a novel propagation model for the smart deployment of MCS. The proposed system’s main contribution is to develop an intelligent environment that eliminates limitations and minimizes related signal fading prediction errors. In addition, having better connectivity and efficiency will resolve the communication problem of smart cities. The chapter also provides a case study that significantly outperforms other empirical models with an accuracy of 95.4%.
- 2 February 2022
- Published by IntechOpen
Open Access3D Point Cloud-Based Tree Canopy Visualization for a Smart Deployment of Mobile Communication Systems
Mobile communication is one of the most important parameters of smart cities in terms of maintaining connectivity and interaction between humans and smart systems. However, In the deployment process of Mobile Communication Systems (MCS), Radio Frequency (RF) engineers use location depended empirical Signal Strength Path Loss (SSPL) models ending up with poor signal strength and slow data connection. This is due to the fact that empirical propagation models usually are restrained by the environment and do not implement state of the art technologies, including Unmanned Aerial Vehicles (UAV), Light Detection and Ranging (LiDAR), Image Processing, and Machine Learning to increase efficiency. Terrains involving buildings, hills, trees, mountains, and human-made structures are considered irregular terrains by telecommunication engineers. Irregular terrains, specifically trees, significantly affect MCS’s efficiency because of their complex pattern resulting in erroneous signal fading via multi-path reflection and absorption. Therefore, a virtual 3D environment is required to extract the required 3D terrain pattern and elevation data from the environment. Once this data is processed in the machine learning algorithm, an adaptive propagation model can be formed and can significantly improve SSPL prediction accuracy for MCS. This chapter presents 3D point cloud visualization via sensor fusion and 2D image color classification techniques, which lead to a novel propagation model for the smart deployment of MCS. The proposed system’s main contribution is to develop an intelligent environment that eliminates limitations and minimizes related signal fading prediction errors. In addition, having better connectivity and efficiency will resolve the communication problem of smart cities. The chapter also provides a case study that significantly outperforms other empirical models with an accuracy of 95.4%.
- Open Access
- 1 February 2022
- Published by Walter de Gruyter GmbH in Journal of Electrical Engineering
- Vol. 73 (1) , 19-27
Basketball self training shooting posture recognition and trajectory estimation using computer vision and Kalman filter
Self-shooting training is one of the fundamental criteria for success in basketball. Particularly, young players increase their performance with regular training. However, the training process becomes painful and time-consuming without a coach since the incorrect shooting posture causes missing shots, leading to reluctance. In this research, a self-shooting posture algorithm is developed to track the movement of basketball players and give them feedback about their position, angle, and basketball projectile trajectory information. The proposed algorithm uses computer vision techniques and Kalman filter to detect the best projectile trajectory using initial conditions such as acceleration due to gravity the initial velocity at the angle of launch having certain horizontal distance to the rim and the rim distance from the ground The acceleration of both gravity and air drag are altered by predefined parameters, including the drag coefficient basketball mass ball radius and silhouette area The proposed algorithm provides the shooting angle in real-time by placing the projectile angle on to the cropped image of the player posture and draws the projectile trajectory towards the basketball hoop According to the results, the players having a specified height can achieve the best shooting at the angle with air drag force. On the other hand, if there is no air resistance, the best shooting angle is deviated significantly. The other stats that are a total time of travel, maximum horizontal distance, maximum height and the time until the top are also given along with the results.
- 1 February 2022
- Published by Walter de Gruyter GmbH in Journal of Electrical Engineering
- Vol. 73 (1) , 19-27
Open AccessBasketball self training shooting posture recognition and trajectory estimation using computer vision and Kalman filter
Self-shooting training is one of the fundamental criteria for success in basketball. Particularly, young players increase their performance with regular training. However, the training process becomes painful and time-consuming without a coach since the incorrect shooting posture causes missing shots, leading to reluctance. In this research, a self-shooting posture algorithm is developed to track the movement of basketball players and give them feedback about their position, angle, and basketball projectile trajectory information. The proposed algorithm uses computer vision techniques and Kalman filter to detect the best projectile trajectory using initial conditions such as acceleration due to gravity the initial velocity at the angle of launch having certain horizontal distance to the rim and the rim distance from the ground The acceleration of both gravity and air drag are altered by predefined parameters, including the drag coefficient basketball mass ball radius and silhouette area The proposed algorithm provides the shooting angle in real-time by placing the projectile angle on to the cropped image of the player posture and draws the projectile trajectory towards the basketball hoop According to the results, the players having a specified height can achieve the best shooting at the angle with air drag force. On the other hand, if there is no air resistance, the best shooting angle is deviated significantly. The other stats that are a total time of travel, maximum horizontal distance, maximum height and the time until the top are also given along with the results.
- 1 April 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Modeling Tree Canopy Signal Power Path Loss (SPPL) for Deployment of Wireless Communication Systems (WCS) Using Point Cloud and Sensor Fusion
Deployment of WCS in different areas requires an expensive and tedious process in terms of obtaining reliable coverage and connectivity. The current random deployment strategies still face some challenges that are related to the complex structure of the environment such as areas covered with tree canopies and hills. These type of structures do not only cause unpredictable shadowing effects through scattering and absorption but also cause faulty tolerances and short network lifespans. Furthermore, the current empirical deployment approaches depended on predefined parameters like path loss exponents and random variable $X_{\sigma }$ with $N(\sigma,~\mu =0)$ which are available only for a limited number of environments. In other words, the empirical approaches are not versatile enough to be used for any environment. In this paper, we propose a solution for the challenges caused by tree canopies and hills using point cloud and computer vision techniques. We fuse Light Detection and Ranging (LiDAR) data with the satellite imagery to acquire information about the location, separation, and the size of the tree canopies. The tree canopies on the line of sight direction (LOS) are determined by the local maximum method and median filter. As a result, we achieved a significant improvement over the existing empirical methods in approximating the SPPL.- 1 April 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Modeling Tree Canopy Signal Power Path Loss (SPPL) for Deployment of Wireless Communication Systems (WCS) Using Point Cloud and Sensor Fusion
Deployment of WCS in different areas requires an expensive and tedious process in terms of obtaining reliable coverage and connectivity. The current random deployment strategies still face some challenges that are related to the complex structure of the environment such as areas covered with tree canopies and hills. These type of structures do not only cause unpredictable shadowing effects through scattering and absorption but also cause faulty tolerances and short network lifespans. Furthermore, the current empirical deployment approaches depended on predefined parameters like path loss exponents and random variable $X_{\sigma }$ with $N(\sigma,~\mu =0)$ which are available only for a limited number of environments. In other words, the empirical approaches are not versatile enough to be used for any environment. In this paper, we propose a solution for the challenges caused by tree canopies and hills using point cloud and computer vision techniques. We fuse Light Detection and Ranging (LiDAR) data with the satellite imagery to acquire information about the location, separation, and the size of the tree canopies. The tree canopies on the line of sight direction (LOS) are determined by the local maximum method and median filter. As a result, we achieved a significant improvement over the existing empirical methods in approximating the SPPL.