Ahmet Küçüker
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Ahmet Küçüker has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
- Burhan Barakli
Sakarya University
3 shared publications - Gökay Bayrak
Bursa Technical University
3 shared publications - Mehmet Bayrak
Department of Electrical and Electronics Engineering, Institute of Natural Sciences, Sakarya University, 54050 Sakarya, Türkiye
3 shared publications Alper Yılmaz
2 shared publicationsFevzeddin Ülker
Department of Electrical and Electronics Engineering, Faculty of Engineering, Sakarya University, Sakarya, Turkey
2 shared publicationsBurhan Baraklı
1 shared publication
Latest Publications
- 1 January 2026
- Published by Tech Science Press in Computer Modeling in Engineering & Sciences
- Vol. 146 (2) , 1-10
High-Performance Segmentation of Power Lines in Aerial Images Using a Wavelet-Guided Hybrid Transformer Network
Inspections of power transmission lines (PTLs) conducted using unmanned aerial vehicles (UAVs) are complicated by the fine structure of the lines and complex backgrounds, making accurate and efficient segmentation challenging. This study presents the Wavelet-Guided Transformer U-Net (WGT-UNet) model, a new hybrid network that combines Convolutional Neural Networks (CNNs), Discrete Wavelet Transform (DWT), and Transformer architectures. The model’s primary contribution is based on spatial and channel attention mechanisms derived from wavelet subbands to guide the Transformer’s self-attention structure. Thus, low and high frequency components are separated at each stage using DWT, suppressing structural noise and making linear objects more prominent. The developed design is supported by multi-component hybrid cost functions that simultaneously solve class imbalance, edge sharpness, structural integrity, and spatial regularity issues. Furthermore, high segmentation success has been achieved in producing sharp boundaries and continuous line structures with the DWT-guided attention mechanism. Experiments conducted on the TTPLA dataset reveal that the version using the ConvNeXt backbone outperforms the current state-of-the-art approaches with an F1-Score of 79.33% and an Intersection over Union (IoU) value of 68.38%. The models and visual outputs of the developed method and all compared models can be accessed at https://github.com/burhanbarakli/WGT-UNET.
- 1 January 2026
- Published by Tech Science Press in Computer Modeling in Engineering & Sciences
- Vol. 146 (2) , 1-10
High-Performance Segmentation of Power Lines in Aerial Images Using a Wavelet-Guided Hybrid Transformer Network
Inspections of power transmission lines (PTLs) conducted using unmanned aerial vehicles (UAVs) are complicated by the fine structure of the lines and complex backgrounds, making accurate and efficient segmentation challenging. This study presents the Wavelet-Guided Transformer U-Net (WGT-UNet) model, a new hybrid network that combines Convolutional Neural Networks (CNNs), Discrete Wavelet Transform (DWT), and Transformer architectures. The model’s primary contribution is based on spatial and channel attention mechanisms derived from wavelet subbands to guide the Transformer’s self-attention structure. Thus, low and high frequency components are separated at each stage using DWT, suppressing structural noise and making linear objects more prominent. The developed design is supported by multi-component hybrid cost functions that simultaneously solve class imbalance, edge sharpness, structural integrity, and spatial regularity issues. Furthermore, high segmentation success has been achieved in producing sharp boundaries and continuous line structures with the DWT-guided attention mechanism. Experiments conducted on the TTPLA dataset reveal that the version using the ConvNeXt backbone outperforms the current state-of-the-art approaches with an F1-Score of 79.33% and an Intersection over Union (IoU) value of 68.38%. The models and visual outputs of the developed method and all compared models can be accessed at https://github.com/burhanbarakli/WGT-UNET.
- Open Access
- 29 December 2025
- Published by Sakarya University Journal of Computer and Information Sciences in Sakarya University Journal of Computer and Information Sciences
- Vol. 8 (4) , 798-811
TriaNet: A Tri-Fusion Attention Network for Segmenting Polyps with Ambiguous Boundaries
Colorectal cancer (CRC) is one of the most common and deadly types of cancer worldwide. During standard colonoscopy procedures to detect polyps, which are early-stage precancerous lesions critical for disease prevention, challenges exist, such as overlooking polyps and the inability to accurately segment polyps with weak borders that are integrated with surrounding tissue using current computer-aided methods. This study proposes a new deep learning architecture, called TriaNet (Tri-Fusion Attention Network), to enhance the segmentation accuracy of polyps with weak borders. The fundamental innovation of TriaNet is its unique “triple-fusion” attention mechanism, which combines three complementary information streams. The proposed method dynamically fuses edge feature information obtained from a hybrid block containing Scharr, DoG, and Gabor filters, the semantic feature map from the decoder structure, and an instantaneous boundary map derived from a Scharr operator applied to an upper layer prediction. Furthermore, Deformable Alignment layers are utilized in skip connections to enhance the model's ability to adapt to variable polyp morphologies. The TriaNET architecture was tested on four different benchmark datasets, including Kvasir-SEG, CVC-ColonDB, ETIS-LaribPolypDB, and CVC-300, which demonstrated superior performance compared to state-of-the-art methods.
- 29 December 2025
- Published by Sakarya University Journal of Computer and Information Sciences in Sakarya University Journal of Computer and Information Sciences
- Vol. 8 (4) , 798-811
Open AccessTriaNet: A Tri-Fusion Attention Network for Segmenting Polyps with Ambiguous Boundaries
Colorectal cancer (CRC) is one of the most common and deadly types of cancer worldwide. During standard colonoscopy procedures to detect polyps, which are early-stage precancerous lesions critical for disease prevention, challenges exist, such as overlooking polyps and the inability to accurately segment polyps with weak borders that are integrated with surrounding tissue using current computer-aided methods. This study proposes a new deep learning architecture, called TriaNet (Tri-Fusion Attention Network), to enhance the segmentation accuracy of polyps with weak borders. The fundamental innovation of TriaNet is its unique “triple-fusion” attention mechanism, which combines three complementary information streams. The proposed method dynamically fuses edge feature information obtained from a hybrid block containing Scharr, DoG, and Gabor filters, the semantic feature map from the decoder structure, and an instantaneous boundary map derived from a Scharr operator applied to an upper layer prediction. Furthermore, Deformable Alignment layers are utilized in skip connections to enhance the model's ability to adapt to variable polyp morphologies. The TriaNET architecture was tested on four different benchmark datasets, including Kvasir-SEG, CVC-ColonDB, ETIS-LaribPolypDB, and CVC-300, which demonstrated superior performance compared to state-of-the-art methods.
- 6 February 2024
- Published by Springer Nature in Electrical Engineering
- Vol. 106 (4) , 4647-4663
A combined covariance analysis and fuzzy logic controller-based protection methodology for discrimination of disturbance and asymmetrical/symmetrical events
- 6 February 2024
- Published by Springer Nature in Electrical Engineering
- Vol. 106 (4) , 4647-4663
A combined covariance analysis and fuzzy logic controller-based protection methodology for discrimination of disturbance and asymmetrical/symmetrical events
- Open Access
- 27 October 2022
- Published by Springer Nature in Research Square
A Combined Covariance Analysis and Fuzzy Logic Controller Based Protection Methodology for Discrimination of Uncertain and Asymmetrical/Symmetrical Events
Distributed energy resources (DERs) such as wind turbines and solar panels have fluctuating power generation profile. Also, uncertainty of DERs & loads can bring about complexity in functioning of protection for Microgrids (MGs). Thus, the operation of classical protection methods used in distribution networks may fail for MGs. In this context, this study presents a new protection methodology for discriminating between uncertain and asymmetrical / symmetrical events in MG. A detection method using covariance analysis of wavelet transform (CAWT) and a classification method based on fuzzy logic controller (FLC) using this analysis are presented. The proposed methods are evaluated under various conditions, including fault resistance, fault inception angle, and fault location. Furthermore, variation of solar irradiance levels and source strength has been taken into consideration to validate the performance of the proposed protection methodology. An advantage of the proposed methodology is that the relationship among phases using a covariance analysis proves the presence of the fault. Another advantage is that the discrimination between uncertain and ASF&SF events was discussed, and the faulted phase(s) was/were accurately identified through FLC. Besides, the efficiency of introduced CAWT&FLC-based methods was confirmed on MG configurations such as radial and mesh. In addition, the presented protection methodology became more efficient when combined with the detection of uncertain events.
- 27 October 2022
- Published by Springer Nature in Research Square
Open AccessA Combined Covariance Analysis and Fuzzy Logic Controller Based Protection Methodology for Discrimination of Uncertain and Asymmetrical/Symmetrical Events
Distributed energy resources (DERs) such as wind turbines and solar panels have fluctuating power generation profile. Also, uncertainty of DERs & loads can bring about complexity in functioning of protection for Microgrids (MGs). Thus, the operation of classical protection methods used in distribution networks may fail for MGs. In this context, this study presents a new protection methodology for discriminating between uncertain and asymmetrical / symmetrical events in MG. A detection method using covariance analysis of wavelet transform (CAWT) and a classification method based on fuzzy logic controller (FLC) using this analysis are presented. The proposed methods are evaluated under various conditions, including fault resistance, fault inception angle, and fault location. Furthermore, variation of solar irradiance levels and source strength has been taken into consideration to validate the performance of the proposed protection methodology. An advantage of the proposed methodology is that the relationship among phases using a covariance analysis proves the presence of the fault. Another advantage is that the discrimination between uncertain and ASF&SF events was discussed, and the faulted phase(s) was/were accurately identified through FLC. Besides, the efficiency of introduced CAWT&FLC-based methods was confirmed on MG configurations such as radial and mesh. In addition, the presented protection methodology became more efficient when combined with the detection of uncertain events.
- 9 June 2022
- Published by Elsevier in International Journal of Hydrogen Energy
- Vol. 48 (18) , 6824-6836
Deep learning-based multi-model ensemble method for classification of PQDs in a hydrogen energy-based microgrid using modified weighted majority algorithm
- 9 June 2022
- Published by Elsevier in International Journal of Hydrogen Energy
- Vol. 48 (18) , 6824-6836
Deep learning-based multi-model ensemble method for classification of PQDs in a hydrogen energy-based microgrid using modified weighted majority algorithm
- 26 February 2022
- Published by Elsevier in International Journal of Hydrogen Energy
- Vol. 47 (45) , 19797-19809
Automated classification of power quality disturbances in a SOFC&PV-based distributed generator using a hybrid machine learning method with high noise immunity
- 26 February 2022
- Published by Elsevier in International Journal of Hydrogen Energy
- Vol. 47 (45) , 19797-19809
Automated classification of power quality disturbances in a SOFC&PV-based distributed generator using a hybrid machine learning method with high noise immunity
- 29 May 2021
- Published by Duzce Universitesi Bilim ve Teknoloji Dergisi in Düzce Üniversitesi Bilim ve Teknoloji Dergisi
- Vol. 9 (3) , 430-445
Karar Destek Makineleri ve Rastgele Orman Ağaçları Yöntemleri ile Vücut Yağ Yüzdesinin Tahmini
Obezite, önemli bir sağlık problemidir ve yaygınlığı giderek artmaktadır. Bu hastalığın, diğer birçok hastalığın tetikleyicisi ve habercisi olduğu bilinmektedir. Obezite hastalığının tedavi sürecinden önce, vücut yağ yüzdesinin doğru bir şekilde tespit edilmesi önemlidir. Yüksek maliyetli yöntemler ile vücut yağ yüzdesi kesin olarak ölçülmektedir. Bu çalışmada, kişilerden alınan antropometrik veri seti ile vücut yağı yüzdesi tespitinin doğru ve maliyetsiz bir şekilde tahmin edilebilmesi için destek vektör regresyonu ile rastgele orman ağaçları regresyon yöntemleri uygulanmıştır. Regresyon yöntemlerinde, model parametre değerleri, veri sayısı, özellik sayısı ve özellik seçimi tahmin başarımında önemlidir. 13 özellikli vücut yağ yüzdesi veri seti baz olarak alınan çalışmada, literatürde sıklıkla kullanılan 25 istatiski yöntem (çarpıklık, merkezi moment, basıklık vb.) ile yeni bir veri seti oluşturulmuş ve elde edilen yeni veri setinin başarım performansının literatürde yer alan diğer çalışmalardan yüksek olduğu görülmüştür. Regresyon parametrelerinin ızgara tarama yöntemleriyle belirlenmesi sayesinde tahmin doğrulukları arttırılmıştır. Ayrıca özellik azaltma yöntemleri ile vücut yağ yüzdesi ile yüksek ilintili özellikler belirlenmiştir. Seçilen özellikler ile gerçekleştirilen regresyon yöntemlerinin tahmin başarı performansının da benzer diğer çalışmalardan yüksek olduğu gözlenmiştir. En iyi ortalama karesel hata değerleri olarak, Rasgele Orman Ağaçları Yöntemi ve istatistiki yöntemle oluşturulan yeni veri seti ile gerçekleştirilen deneyde 2,2519 değeri elde edilirken, Karar Destek Makinaları ve en iyi 6 F-skor değerine sahip özellikler ile yapılan regresyon deneyinde 3,174 değerine ulaşılmıştır.
- 29 May 2021
- Published by Duzce Universitesi Bilim ve Teknoloji Dergisi in Düzce Üniversitesi Bilim ve Teknoloji Dergisi
- Vol. 9 (3) , 430-445
Karar Destek Makineleri ve Rastgele Orman Ağaçları Yöntemleri ile Vücut Yağ Yüzdesinin Tahmini
Obezite, önemli bir sağlık problemidir ve yaygınlığı giderek artmaktadır. Bu hastalığın, diğer birçok hastalığın tetikleyicisi ve habercisi olduğu bilinmektedir. Obezite hastalığının tedavi sürecinden önce, vücut yağ yüzdesinin doğru bir şekilde tespit edilmesi önemlidir. Yüksek maliyetli yöntemler ile vücut yağ yüzdesi kesin olarak ölçülmektedir. Bu çalışmada, kişilerden alınan antropometrik veri seti ile vücut yağı yüzdesi tespitinin doğru ve maliyetsiz bir şekilde tahmin edilebilmesi için destek vektör regresyonu ile rastgele orman ağaçları regresyon yöntemleri uygulanmıştır. Regresyon yöntemlerinde, model parametre değerleri, veri sayısı, özellik sayısı ve özellik seçimi tahmin başarımında önemlidir. 13 özellikli vücut yağ yüzdesi veri seti baz olarak alınan çalışmada, literatürde sıklıkla kullanılan 25 istatiski yöntem (çarpıklık, merkezi moment, basıklık vb.) ile yeni bir veri seti oluşturulmuş ve elde edilen yeni veri setinin başarım performansının literatürde yer alan diğer çalışmalardan yüksek olduğu görülmüştür. Regresyon parametrelerinin ızgara tarama yöntemleriyle belirlenmesi sayesinde tahmin doğrulukları arttırılmıştır. Ayrıca özellik azaltma yöntemleri ile vücut yağ yüzdesi ile yüksek ilintili özellikler belirlenmiştir. Seçilen özellikler ile gerçekleştirilen regresyon yöntemlerinin tahmin başarı performansının da benzer diğer çalışmalardan yüksek olduğu gözlenmiştir. En iyi ortalama karesel hata değerleri olarak, Rasgele Orman Ağaçları Yöntemi ve istatistiki yöntemle oluşturulan yeni veri seti ile gerçekleştirilen deneyde 2,2519 değeri elde edilirken, Karar Destek Makinaları ve en iyi 6 F-skor değerine sahip özellikler ile yapılan regresyon deneyinde 3,174 değerine ulaşılmıştır.
- 1 June 2019
- Published by Sakarya University Journal of Science in Sakarya University Journal of Science
- Vol. 23 (3) , 413-424
Traveling Wave and Wavelet Based Fault Location Detection in Microgrids
Demand for renewable energy sources is increasing day by day. With the increase in demand for renewable energy sources, the use of microgrids also increases. Microgrids, compared to traditional grids, have a more complex structure and need more complex protection functions. In this study, a traveling wave-based fault location detection system for microgrids with loop distribution systems (looped microgrids) was developed. In traveling wave-based fault location, measurements taken from one or both ends of a line are used. However, in looped grids, lines do not have a beginning or an end. In order to allow the use of traveling wave-based protection in loop distribution systems, the system is separated into three zones. To each three separate zone a type D fault locator method was applied, and by ensuring that these algorithms worked in accordance, fault location detection was carried out in a looped microgrid.
- 1 June 2019
- Published by Sakarya University Journal of Science in Sakarya University Journal of Science
- Vol. 23 (3) , 413-424
Traveling Wave and Wavelet Based Fault Location Detection in Microgrids
Demand for renewable energy sources is increasing day by day. With the increase in demand for renewable energy sources, the use of microgrids also increases. Microgrids, compared to traditional grids, have a more complex structure and need more complex protection functions. In this study, a traveling wave-based fault location detection system for microgrids with loop distribution systems (looped microgrids) was developed. In traveling wave-based fault location, measurements taken from one or both ends of a line are used. However, in looped grids, lines do not have a beginning or an end. In order to allow the use of traveling wave-based protection in loop distribution systems, the system is separated into three zones. To each three separate zone a type D fault locator method was applied, and by ensuring that these algorithms worked in accordance, fault location detection was carried out in a looped microgrid.
- 1 March 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Design and control of photovoltaic/wind/battery based microgrid system
This paper deals with the design and control of a micro-grid, including various alternative energy resources (photovoltaic and wind) and battery energy storage system which operates in stand-alone as well as in grid-connected mode. The proposed micro-grid is controlled via various non-isolated converters while an energy management is performed through switching based algorithm. According to the strategy, the wind is used as the primary power source while the Photovoltaic (PV) is added to improve the reliability of system under different weather conditions. The battery module is utilized as an energy storage system during surplus power and/or backup device during demand. The proposed system used real record of weather pattern and load conditions for a small community at Islamabad, Pakistan. This city is gifted with several natural resources that can generate a significant amount of power energy for the region. MATLAB simulation results show the effectiveness of proposed system in terms of grid stability, power sharing, load tracking and power quality.
- 1 March 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Design and control of photovoltaic/wind/battery based microgrid system
This paper deals with the design and control of a micro-grid, including various alternative energy resources (photovoltaic and wind) and battery energy storage system which operates in stand-alone as well as in grid-connected mode. The proposed micro-grid is controlled via various non-isolated converters while an energy management is performed through switching based algorithm. According to the strategy, the wind is used as the primary power source while the Photovoltaic (PV) is added to improve the reliability of system under different weather conditions. The battery module is utilized as an energy storage system during surplus power and/or backup device during demand. The proposed system used real record of weather pattern and load conditions for a small community at Islamabad, Pakistan. This city is gifted with several natural resources that can generate a significant amount of power energy for the region. MATLAB simulation results show the effectiveness of proposed system in terms of grid stability, power sharing, load tracking and power quality.