Burhan Ergen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Burhan Ergen has a total of 38 co-authors. Here are the top co-authors by number of shared publications:
- Abdullah Şener
Management Information Systems, Faculty of Economics and Administrative Sciences, Fırat University, Elazığ, Turkey
17 shared publications - Mesut Toğaçar
Department of Management Information Systems, Faculty of Economics and Administrative Sciences, Fırat University, 23119 Elazığ, Türkiye
14 shared publications - Gürkan Doğan
Munzur Üniversitesi
11 shared publications - Zafer Cömert
Samsun Üniversitesi
10 shared publications - Meryem Durmuş
Distance Education Center, Rectorate, Samsun University, Samsun, Trkiye
4 shared publications - Abdulkadir Sengur
Department of Electrical-Electronic Engineering, Faculty of Technology, Firat University, 23000 Elazig, Turkey
3 shared publications
Latest Publications
- Open Access
- 29 September 2026
- Published by MDPI AG in Electronics
- Vol. 15 (19) , 4470
A Hybrid Deep Learning Approach for Early Detection of Wildfires Using Transformer Models and Feature Fusion Techniques
Natural disasters, particularly wildfires, pose a serious threat to ecosystems, human lives, and infrastructure, making rapid and accurate fire detection essential for effective disaster management. Despite recent advances in deep learning-based wildfire detection, conventional CNN-based approaches are inherently limited in modeling long-range spatial dependencies, while relying on a single deep architecture may provide an incomplete representation of complex and multi-scale fire patterns. Furthermore, high-dimensional deep features may contain redundant information, increasing computational complexity and limiting the efficiency of real-time monitoring systems. Therefore, there remains a need for an efficient framework that can exploit complementary global representations from multiple transformer architectures while retaining only the most discriminative features. To investigate this application-specific problem, this study evaluates a hybrid framework that integrates three heterogeneous Vision Transformer architectures—DeiT3, MaxViT, and Swin—with feature-level fusion, mRMR-based feature selection, and SVM classification for binary wildfire image classification. The publicly available FlameVision dataset, consisting of aerial images categorized into fire and non-fire classes, was used for evaluation. Deep feature representations extracted from the three transformer architectures were systematically fused to exploit their complementary characteristics. Minimum Redundancy Maximum Relevance (mRMR) feature selection was subsequently employed to reduce feature redundancy and retain the most discriminative information, followed by classification using Support Vector Machines (SVM). The experimental results demonstrated an overall classification accuracy of up to 100% on the internal hold-out test partition and in five-fold cross-validation within the FlameVision dataset. Because both evaluation procedures were conducted using images originating from the same source dataset, these results should be interpreted as within-dataset classification performance rather than evidence of external generalization. Validation on independently collected wildfire datasets is therefore required to determine the robustness of the proposed framework under different geographical, sensor, environmental, and acquisition conditions. The proposed framework demonstrates that complementary transformer-derived representations combined with feature selection can provide highly discriminative wildfire representations while substantially reducing the number of features required for classification. These findings demonstrate that the proposed approach can retain high classification performance using a substantially more compact feature representation. However, the resulting reduction in feature dimensionality should not be interpreted as evidence of improved runtime efficiency, since computational time before and after feature selection was not experimentally measured. However, real-time performance is not claimed in the present study because FPS, end-to-end latency, runtime memory consumption, and deployment on continuous UAV or satellite streams were not experimentally evaluated. Such operational evaluation is left for future work.
- 29 September 2026
- Published by MDPI AG in Electronics
- Vol. 15 (19) , 4470
Open AccessA Hybrid Deep Learning Approach for Early Detection of Wildfires Using Transformer Models and Feature Fusion Techniques
Natural disasters, particularly wildfires, pose a serious threat to ecosystems, human lives, and infrastructure, making rapid and accurate fire detection essential for effective disaster management. Despite recent advances in deep learning-based wildfire detection, conventional CNN-based approaches are inherently limited in modeling long-range spatial dependencies, while relying on a single deep architecture may provide an incomplete representation of complex and multi-scale fire patterns. Furthermore, high-dimensional deep features may contain redundant information, increasing computational complexity and limiting the efficiency of real-time monitoring systems. Therefore, there remains a need for an efficient framework that can exploit complementary global representations from multiple transformer architectures while retaining only the most discriminative features. To investigate this application-specific problem, this study evaluates a hybrid framework that integrates three heterogeneous Vision Transformer architectures—DeiT3, MaxViT, and Swin—with feature-level fusion, mRMR-based feature selection, and SVM classification for binary wildfire image classification. The publicly available FlameVision dataset, consisting of aerial images categorized into fire and non-fire classes, was used for evaluation. Deep feature representations extracted from the three transformer architectures were systematically fused to exploit their complementary characteristics. Minimum Redundancy Maximum Relevance (mRMR) feature selection was subsequently employed to reduce feature redundancy and retain the most discriminative information, followed by classification using Support Vector Machines (SVM). The experimental results demonstrated an overall classification accuracy of up to 100% on the internal hold-out test partition and in five-fold cross-validation within the FlameVision dataset. Because both evaluation procedures were conducted using images originating from the same source dataset, these results should be interpreted as within-dataset classification performance rather than evidence of external generalization. Validation on independently collected wildfire datasets is therefore required to determine the robustness of the proposed framework under different geographical, sensor, environmental, and acquisition conditions. The proposed framework demonstrates that complementary transformer-derived representations combined with feature selection can provide highly discriminative wildfire representations while substantially reducing the number of features required for classification. These findings demonstrate that the proposed approach can retain high classification performance using a substantially more compact feature representation. However, the resulting reduction in feature dimensionality should not be interpreted as evidence of improved runtime efficiency, since computational time before and after feature selection was not experimentally measured. However, real-time performance is not claimed in the present study because FPS, end-to-end latency, runtime memory consumption, and deployment on continuous UAV or satellite streams were not experimentally evaluated. Such operational evaluation is left for future work.
- Open Access
- 3 September 2026
- Published by Kahramanmaras Sutcu Imam University Journal of Engineering Sciences in Kahramanmaraş Sütçü İmam Üniversitesi Mühendislik Bilimleri Dergisi
- Vol. 29 (3) , 1087-1100
AFET SONRASI YENİDEN YAPILAŞMA SÜREÇLERİNDE HIZLANDIRICI BİR ARAÇ OLARAK YAPAY ZEKA DESTEKLİ OTOMATİK KAT PLANI ÜRETİMİ
Büyük ölçekli afetler sonrasında ortaya çıkan acil konut ihtiyacı, kısa süre içinde binlerce konut biriminin hızlı ve nitelikli biçimde tasarlanmasını zorunlu kılmaktadır. Geleneksel mimari tasarım yöntemleri, her bir konut biriminin ayrı ayrı ele alınmasını gerektirdiğinden, zaman ve iş gücü bakımından ölçeklenebilirlik açısından sınırlı kalmaktadır. Bu çalışmada, oda sayısı, oda türü ve mekânsal bağlantı şeması gibi kullanıcı tarafından tanımlanan sınırlı tasarım kısıtlarından otomatik olarak kat planı üreten hibrit bir öğrenme çerçevesi önerilmektedir. Söz konusu kısıtlar, vaziyet planı ve yapı sınırları bağlamında temsil edilmekte; mekânsal ilişkilerin kodlanması amacıyla İkinci Nesil Grafik Dikkat Ağı (GATv2) mimarisi kullanılmaktadır. GATv2 modülü tarafından üretilen vektörel yerleşim çıktıları, diferansiyellenebilir bir işleme katmanından geçirilerek koşullu üretken çekişmeli ağ (cGAN) aracılığıyla görsel doku ve mobilya bilgisi içeren nihai kat planlarına dönüştürülmektedir. RPLAN veri setinde yer alan 80.000 etiketli plan üzerinde eğitilen ve optimize edilen model, 0,7649 IoU yerleşim doğruluğu ve 22,10 Fréchet Inception Distance (FID) değeri elde etmiştir. Bulgular, afet sonrası toplu konut üretiminde tasarım esnekliğini korurken üretim süresini belirgin biçimde azaltabilecek bir karar destek aracının uygulanabilirliğini ortaya koymaktadır.
- 3 September 2026
- Published by Kahramanmaras Sutcu Imam University Journal of Engineering Sciences in Kahramanmaraş Sütçü İmam Üniversitesi Mühendislik Bilimleri Dergisi
- Vol. 29 (3) , 1087-1100
Open AccessAFET SONRASI YENİDEN YAPILAŞMA SÜREÇLERİNDE HIZLANDIRICI BİR ARAÇ OLARAK YAPAY ZEKA DESTEKLİ OTOMATİK KAT PLANI ÜRETİMİ
Büyük ölçekli afetler sonrasında ortaya çıkan acil konut ihtiyacı, kısa süre içinde binlerce konut biriminin hızlı ve nitelikli biçimde tasarlanmasını zorunlu kılmaktadır. Geleneksel mimari tasarım yöntemleri, her bir konut biriminin ayrı ayrı ele alınmasını gerektirdiğinden, zaman ve iş gücü bakımından ölçeklenebilirlik açısından sınırlı kalmaktadır. Bu çalışmada, oda sayısı, oda türü ve mekânsal bağlantı şeması gibi kullanıcı tarafından tanımlanan sınırlı tasarım kısıtlarından otomatik olarak kat planı üreten hibrit bir öğrenme çerçevesi önerilmektedir. Söz konusu kısıtlar, vaziyet planı ve yapı sınırları bağlamında temsil edilmekte; mekânsal ilişkilerin kodlanması amacıyla İkinci Nesil Grafik Dikkat Ağı (GATv2) mimarisi kullanılmaktadır. GATv2 modülü tarafından üretilen vektörel yerleşim çıktıları, diferansiyellenebilir bir işleme katmanından geçirilerek koşullu üretken çekişmeli ağ (cGAN) aracılığıyla görsel doku ve mobilya bilgisi içeren nihai kat planlarına dönüştürülmektedir. RPLAN veri setinde yer alan 80.000 etiketli plan üzerinde eğitilen ve optimize edilen model, 0,7649 IoU yerleşim doğruluğu ve 22,10 Fréchet Inception Distance (FID) değeri elde etmiştir. Bulgular, afet sonrası toplu konut üretiminde tasarım esnekliğini korurken üretim süresini belirgin biçimde azaltabilecek bir karar destek aracının uygulanabilirliğini ortaya koymaktadır.
- 1 September 2026
- Published by Igdir University in Journal of the Institute of Science and Technology
- Vol. 16 (3) , 904-924
Quantifying Post-Earthquake Urban Green Space Loss Using an Optimized Deep Learning-Based Semantic Segmentation Framework
Large-scale earthquakes and the rapid reconstruction processes that follow lead to significant spatial transformations in urban ecosystems, particularly exerting pronounced impacts on intra-urban green spaces. This study quantitatively evaluates changes in urban green areas that occurred after the 2023 Kahramanmaraş earthquakes in Hatay (Antakya) and Elazığ (Güneykent) by utilizing high-resolution satellite imagery and an optimized deep learning–based semantic segmentation approach. In the proposed framework, an enhanced U-Net architecture incorporating asymmetric and dilated convolutional layers was designed to improve parameter efficiency while preserving extensive contextual information. The proposed model, comprising approximately 1.7 million parameters, achieved 88% accuracy, 0.76 Intersection over Union (IoU), and a 0.86 Dice score, demonstrating competitive performance compared to contemporary segmentation models with substantially higher parameter counts. Pixel-level comparison of pre- and post-earthquake imagery revealed reductions in green space reaching up to 89% in the examined sub-regions, with even higher rates observed in certain localized areas. The findings indicate that intensive post-disaster construction and urban reorganization processes may impose substantial pressure on environmental sustainability. The proposed method provides a computationally efficient and scalable decision-support framework for monitoring post-disaster environmental changes and developing sustainable urban planning strategies.
- 1 September 2026
- Published by Igdir University in Journal of the Institute of Science and Technology
- Vol. 16 (3) , 904-924
Quantifying Post-Earthquake Urban Green Space Loss Using an Optimized Deep Learning-Based Semantic Segmentation Framework
Large-scale earthquakes and the rapid reconstruction processes that follow lead to significant spatial transformations in urban ecosystems, particularly exerting pronounced impacts on intra-urban green spaces. This study quantitatively evaluates changes in urban green areas that occurred after the 2023 Kahramanmaraş earthquakes in Hatay (Antakya) and Elazığ (Güneykent) by utilizing high-resolution satellite imagery and an optimized deep learning–based semantic segmentation approach. In the proposed framework, an enhanced U-Net architecture incorporating asymmetric and dilated convolutional layers was designed to improve parameter efficiency while preserving extensive contextual information. The proposed model, comprising approximately 1.7 million parameters, achieved 88% accuracy, 0.76 Intersection over Union (IoU), and a 0.86 Dice score, demonstrating competitive performance compared to contemporary segmentation models with substantially higher parameter counts. Pixel-level comparison of pre- and post-earthquake imagery revealed reductions in green space reaching up to 89% in the examined sub-regions, with even higher rates observed in certain localized areas. The findings indicate that intensive post-disaster construction and urban reorganization processes may impose substantial pressure on environmental sustainability. The proposed method provides a computationally efficient and scalable decision-support framework for monitoring post-disaster environmental changes and developing sustainable urban planning strategies.
- Open Access
- 17 August 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
CalFireSegNet: a lightweight hybrid attention–transformer network for post-wildfire building footprint change analysis using high-resolution satellite imagery
Natural disasters, particularly wildfires, cause severe human, environmental, and economic losses worldwide. Rapid and accurate identification of building footprints and potential potential structural changes is essential for effective emergency response, search-and-rescue operations, and post-disaster recovery planning. To address the challenges of identifying building loss from remote sensing imagery, this study proposes CalFireSegNet, a lightweight hybrid attention–transformer network for post-wildfire building footprint extraction and loss proxy detection from satellite imagery. The proposed architecture integrates depthwise convolutions, convolutional block attention modules (CBAM), atrous spatial pyramid pooling (ASPP), and Transformer blocks to effectively capture both local structural details and long-range contextual dependencies while maintaining low computational complexity. The model was trained and evaluated using benchmark building segmentation datasets (Inria and WHU) and subsequently applied to pre- and post-event satellite imagery from the recent California wildfires. Experimental results demonstrate that CalFireSegNet achieves superior performance compared with several state-of-the-art semantic segmentation models, including U-Net, PSPNet, DeepLabv3+, ENet, HRNet, and SegNet, obtaining 98.45% accuracy, 94.35% mIoU, and 95.03% Dice Similarity Score while requiring only 3.72 million parameters. Furthermore, a lightweight mask-difference framework was developed to generate a spatial proxy of potential building footprint loss using pre- and post-event satellite pairs. Since publicly available building-level damage annotations for recent California wildfire events remain limited, the real-world wildfire experiments are presented as a validation of cross-domain applicability rather than a fully supervised structural loss proxy estimation benchmark.
- 17 August 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
Open AccessCalFireSegNet: a lightweight hybrid attention–transformer network for post-wildfire building footprint change analysis using high-resolution satellite imagery
Natural disasters, particularly wildfires, cause severe human, environmental, and economic losses worldwide. Rapid and accurate identification of building footprints and potential potential structural changes is essential for effective emergency response, search-and-rescue operations, and post-disaster recovery planning. To address the challenges of identifying building loss from remote sensing imagery, this study proposes CalFireSegNet, a lightweight hybrid attention–transformer network for post-wildfire building footprint extraction and loss proxy detection from satellite imagery. The proposed architecture integrates depthwise convolutions, convolutional block attention modules (CBAM), atrous spatial pyramid pooling (ASPP), and Transformer blocks to effectively capture both local structural details and long-range contextual dependencies while maintaining low computational complexity. The model was trained and evaluated using benchmark building segmentation datasets (Inria and WHU) and subsequently applied to pre- and post-event satellite imagery from the recent California wildfires. Experimental results demonstrate that CalFireSegNet achieves superior performance compared with several state-of-the-art semantic segmentation models, including U-Net, PSPNet, DeepLabv3+, ENet, HRNet, and SegNet, obtaining 98.45% accuracy, 94.35% mIoU, and 95.03% Dice Similarity Score while requiring only 3.72 million parameters. Furthermore, a lightweight mask-difference framework was developed to generate a spatial proxy of potential building footprint loss using pre- and post-event satellite pairs. Since publicly available building-level damage annotations for recent California wildfire events remain limited, the real-world wildfire experiments are presented as a validation of cross-domain applicability rather than a fully supervised structural loss proxy estimation benchmark.
- Open Access
- 15 March 2026
- Published by Politeknik Dergisi in Journal of Polytechnic
- Vol. 29 (2) , 1-10
Hibrit ResNeXt ve LSTM Mimarisi Kullanılarak Deepfake Video Algılama
Deepfake içeriklerin artan yaygınlığı, bireysel gizlilik, medya güvenilirliği ve kamu güveni için ciddi bir tehdit oluşturmaktadır. Mevcut tespit yöntemleri genellikle çeşitli manipülasyon teknikleri ve video kalite seviyeleri arasında genelleme yapmakta zorlanmaktadır. Bu çalışma, ResNeXt-50'nin mekansal özellik çıkarma güçlerini LSTM ağlarının zamansal dizi modelleme yetenekleriyle birlikte kullanan derin öğrenmeye dayalı bir hibrit mimari sunmaktadır. Önerilen çerçeve, başlangıçta önceden eğitilmiş bir ResNeXt-50 omurgası aracılığıyla kare bazlı özellikler elde ederek ve ardından bir LSTM katmanı aracılığıyla zamansal dinamikleri inceleyerek video girişini işler. Deneysel değerlendirmeler, DFDC, Celeb-DF, FaceForensics++ ve DFD dahil olmak üzere kıyaslama veri kümeleri kullanılarak yürütülmüştür. Bulgular, geliştirilen modelin geleneksel CNN-LSTM kombinasyonlarından önemli ölçüde daha iyi performans gösterdiğini, DFDC veri kümesinde %95,7 ve diğer veri kümelerinde %90'ın üzerinde doğruluk elde ettiğini göstermektedir. Bu araştırma, hibrit derin öğrenme tekniklerinin gerçek dünya video kimlik doğrulama sistemlerinde pratik uygulanabilirliğini vurgulamakta ve sentetik ortam tespitinin büyüyen alanına yüksek performanslı bir çözüm sunmaktadır.
- 15 March 2026
- Published by Politeknik Dergisi in Journal of Polytechnic
- Vol. 29 (2) , 1-10
Open AccessHibrit ResNeXt ve LSTM Mimarisi Kullanılarak Deepfake Video Algılama
Deepfake içeriklerin artan yaygınlığı, bireysel gizlilik, medya güvenilirliği ve kamu güveni için ciddi bir tehdit oluşturmaktadır. Mevcut tespit yöntemleri genellikle çeşitli manipülasyon teknikleri ve video kalite seviyeleri arasında genelleme yapmakta zorlanmaktadır. Bu çalışma, ResNeXt-50'nin mekansal özellik çıkarma güçlerini LSTM ağlarının zamansal dizi modelleme yetenekleriyle birlikte kullanan derin öğrenmeye dayalı bir hibrit mimari sunmaktadır. Önerilen çerçeve, başlangıçta önceden eğitilmiş bir ResNeXt-50 omurgası aracılığıyla kare bazlı özellikler elde ederek ve ardından bir LSTM katmanı aracılığıyla zamansal dinamikleri inceleyerek video girişini işler. Deneysel değerlendirmeler, DFDC, Celeb-DF, FaceForensics++ ve DFD dahil olmak üzere kıyaslama veri kümeleri kullanılarak yürütülmüştür. Bulgular, geliştirilen modelin geleneksel CNN-LSTM kombinasyonlarından önemli ölçüde daha iyi performans gösterdiğini, DFDC veri kümesinde %95,7 ve diğer veri kümelerinde %90'ın üzerinde doğruluk elde ettiğini göstermektedir. Bu araştırma, hibrit derin öğrenme tekniklerinin gerçek dünya video kimlik doğrulama sistemlerinde pratik uygulanabilirliğini vurgulamakta ve sentetik ortam tespitinin büyüyen alanına yüksek performanslı bir çözüm sunmaktadır.
- 6 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Comparative Evaluation of Machine Learning Models for Gender Classification from Speech Signals
In this study, a machine learning method was developed to determine the gender of people based on voice signals. For this purpose, spectral, temporal, pitch, MFCC (Mel-Frequency Cepstral Coefficients) and energy-based features of the voice data were extracted and analysed. Different machine learning algorithms were compared to determine the model with the highest hit rate. Experimental analyses showed that the ExtraTreesClassifier model performed the best with an accuracy of 98.81%. The accuracy, loss and confusion matrix analyses showed that the model had a high generalisation ability and did not suffer from overfitting. The error rate was calculated to be 1.56% for the female class and 0.79% for the male class. The results of this study suggest that speech-based gender classification systems could contribute to potential applications in areas such as customer service, medical diagnostics and human-computer interaction.
- 6 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Comparative Evaluation of Machine Learning Models for Gender Classification from Speech Signals
In this study, a machine learning method was developed to determine the gender of people based on voice signals. For this purpose, spectral, temporal, pitch, MFCC (Mel-Frequency Cepstral Coefficients) and energy-based features of the voice data were extracted and analysed. Different machine learning algorithms were compared to determine the model with the highest hit rate. Experimental analyses showed that the ExtraTreesClassifier model performed the best with an accuracy of 98.81%. The accuracy, loss and confusion matrix analyses showed that the model had a high generalisation ability and did not suffer from overfitting. The error rate was calculated to be 1.56% for the female class and 0.79% for the male class. The results of this study suggest that speech-based gender classification systems could contribute to potential applications in areas such as customer service, medical diagnostics and human-computer interaction.
- 28 July 2025
- Published by Springer Nature in Health Information Science and Systems
- Vol. 13 (1) , 1-20
HASK-Net: new hybrid attention selective kernel network for automatic colon cancer detection from colonoscopy images
Colorectal cancer is a frequently analyzed cancer type seen today. It is the third leading of cancer in terms of incidence. It is among the second leading cases resulting in death. However, when detecting pre-cancerous polyps using traditional methods, the rate of being overlooked is high. Therefore, the rate of missed polyps can be greatly reduced with deep learning-based computer-aided diagnosis (CADx) methods. In this study, a novel convolutional neural network based on CADx, called HASK-Net, is proposed for automatic polyp detection. While HASK-Net uses ResNet50 as a feature extractor in the encoder network, the novely designed hybrid attention selective kernel convolution uses it in the decoder network to increase the representation power in feature maps and learn more complex features. In experimental results on the publicly available Hiper-Kvasir and EndoTech 2020 benchmark datasets, HASK-Net showed extremely promising performance with a dice similarity score of 97.02% and 84.83% and an mIoU score of 95.38% and 84.32%, respectively.
- 28 July 2025
- Published by Springer Nature in Health Information Science and Systems
- Vol. 13 (1) , 1-20
HASK-Net: new hybrid attention selective kernel network for automatic colon cancer detection from colonoscopy images
Colorectal cancer is a frequently analyzed cancer type seen today. It is the third leading of cancer in terms of incidence. It is among the second leading cases resulting in death. However, when detecting pre-cancerous polyps using traditional methods, the rate of being overlooked is high. Therefore, the rate of missed polyps can be greatly reduced with deep learning-based computer-aided diagnosis (CADx) methods. In this study, a novel convolutional neural network based on CADx, called HASK-Net, is proposed for automatic polyp detection. While HASK-Net uses ResNet50 as a feature extractor in the encoder network, the novely designed hybrid attention selective kernel convolution uses it in the decoder network to increase the representation power in feature maps and learn more complex features. In experimental results on the publicly available Hiper-Kvasir and EndoTech 2020 benchmark datasets, HASK-Net showed extremely promising performance with a dice similarity score of 97.02% and 84.83% and an mIoU score of 95.38% and 84.32%, respectively.
- Open Access
- 30 June 2025
- Published by Dicle Universitesi Muhendislik Fakultesi Muhendislik Dergisi in Dicle Üniversitesi Mühendislik Fakültesi Mühendislik Dergisi
- Vol. 16 (2) , 357-367
Şüpheli Haberlerin Tespitine Yönelik Derin Öğrenme Tabanlı Haber Teyit Sisteminin Gerçekleştirilmesi
Dijital medyanın yaygınlaşması, sahte haberlerin toplumu olumsuz etkilemesine yol açmaktadır. Özellikle sosyal medya platformları, yanlış bilgilerin hızla yayılmasına sebep olmaktadır. Bu durum bireylerin yanlış yönlendirilmesine, toplumsal kutuplaşmaya ve güven kaybına yol açmaktadır. Bu nedenle, sahte haberlerin hızlı ve güvenilir bir şekilde tespiti çağımızın en önemli sorunlarından biri haline gelmiştir. Bu çalışmada, derin öğrenme ve doğal dil işleme (NLP) tabanlı yenilikçi bir haber teyit sistemi geliştirilmiştir. Sistem, dijital haber içeriklerinin doğruluğunu analiz ederek, kullanıcıları yanlış bilgilendirmeden korumayı hedeflemektedir. Geliştirilen model, haber metinlerini Zemberek kütüphansiyle işleyerek optimize etmekte, ardından BERT tabanlı özetleme yöntemleriyle metinlerin anlamını zenginleştirmektedir. Zemberek, Türkçe metinlerin doğal dil işleme süreçlerini kolaylaştırırken, SBERT modeli ise haber içeriklerinin daha derin anlamsal analizini gerçekleştirmektedir. Ayrıca YAKE algoritması haber içeriklerinden anahtar kelimeler çıkararak güvenilir kaynaklarla karşılaştırılmasını kolaylaştırmaktadır. Metinler arasındaki anlamsal benzerlik SBERT modeli kullanılarak ölçülmektedir. Kelime bazlı benzerlikler ise N-gram yöntemiyle ölçülmektedir. Bu iki yöntemden elde edilen sonuçlar birleştirilerek, haberin doğruluk skoru hesaplanmakta ve kullanıcıya sunulmaktadır. Önerilen sistem, sahte haberlerin tespitinde yüksek doğruluk oranları sunmakta ve gerçek zamanlı veri analizi sayesinde güncel olaylardaki yanlış bilgileri hızla belirlemektedir. Çok dilli ve çok modaliteli bir yapıya kavuşturulması hedeflenen sistemin uluslararası sahte haber tespiti için de etkili bir çözüm sunması hedeflenmektedir.
- 30 June 2025
- Published by Dicle Universitesi Muhendislik Fakultesi Muhendislik Dergisi in Dicle Üniversitesi Mühendislik Fakültesi Mühendislik Dergisi
- Vol. 16 (2) , 357-367
Open AccessŞüpheli Haberlerin Tespitine Yönelik Derin Öğrenme Tabanlı Haber Teyit Sisteminin Gerçekleştirilmesi
Dijital medyanın yaygınlaşması, sahte haberlerin toplumu olumsuz etkilemesine yol açmaktadır. Özellikle sosyal medya platformları, yanlış bilgilerin hızla yayılmasına sebep olmaktadır. Bu durum bireylerin yanlış yönlendirilmesine, toplumsal kutuplaşmaya ve güven kaybına yol açmaktadır. Bu nedenle, sahte haberlerin hızlı ve güvenilir bir şekilde tespiti çağımızın en önemli sorunlarından biri haline gelmiştir. Bu çalışmada, derin öğrenme ve doğal dil işleme (NLP) tabanlı yenilikçi bir haber teyit sistemi geliştirilmiştir. Sistem, dijital haber içeriklerinin doğruluğunu analiz ederek, kullanıcıları yanlış bilgilendirmeden korumayı hedeflemektedir. Geliştirilen model, haber metinlerini Zemberek kütüphansiyle işleyerek optimize etmekte, ardından BERT tabanlı özetleme yöntemleriyle metinlerin anlamını zenginleştirmektedir. Zemberek, Türkçe metinlerin doğal dil işleme süreçlerini kolaylaştırırken, SBERT modeli ise haber içeriklerinin daha derin anlamsal analizini gerçekleştirmektedir. Ayrıca YAKE algoritması haber içeriklerinden anahtar kelimeler çıkararak güvenilir kaynaklarla karşılaştırılmasını kolaylaştırmaktadır. Metinler arasındaki anlamsal benzerlik SBERT modeli kullanılarak ölçülmektedir. Kelime bazlı benzerlikler ise N-gram yöntemiyle ölçülmektedir. Bu iki yöntemden elde edilen sonuçlar birleştirilerek, haberin doğruluk skoru hesaplanmakta ve kullanıcıya sunulmaktadır. Önerilen sistem, sahte haberlerin tespitinde yüksek doğruluk oranları sunmakta ve gerçek zamanlı veri analizi sayesinde güncel olaylardaki yanlış bilgileri hızla belirlemektedir. Çok dilli ve çok modaliteli bir yapıya kavuşturulması hedeflenen sistemin uluslararası sahte haber tespiti için de etkili bir çözüm sunması hedeflenmektedir.
- Open Access
- 27 June 2025
- Published by Bingol Universitesi in Türk Doğa ve Fen Dergisi
- Vol. 14 (2) , 247-258
Enhancing Geometry Education through Deep Learning Models: Addressing Challenges in Three-Dimensional Shape Visualization
Integrating technology into mathematics education is crucial for enhancing the understanding of mathematical concepts and skills, as well as increasing motivation. This is particularly applicable in geometry classes, where technology can facilitate the detection of geometric shapes, impacting both the learning and teaching processes. In this context, emerging concepts of artificial intelligence and deep learning can be utilized as tools to overcome such limitations. This study addresses the challenges that teachers face when drawing three-dimensional geometric shapes in digital environments. Shapes drawn manually in digital environments can often be complex, making it difficult for teachers to create accurate and precise drawings. Deep learning models can assist teachers in correcting drawing errors, thereby providing students with clearer and more comprehensible visuals to facilitate the learning of geometric concepts. The study emphasizes the high accuracy rates achieved using various deep learning models, highlighting their impressive capabilities in accurately classifying geometric shapes.
- 27 June 2025
- Published by Bingol Universitesi in Türk Doğa ve Fen Dergisi
- Vol. 14 (2) , 247-258
Open AccessEnhancing Geometry Education through Deep Learning Models: Addressing Challenges in Three-Dimensional Shape Visualization
Integrating technology into mathematics education is crucial for enhancing the understanding of mathematical concepts and skills, as well as increasing motivation. This is particularly applicable in geometry classes, where technology can facilitate the detection of geometric shapes, impacting both the learning and teaching processes. In this context, emerging concepts of artificial intelligence and deep learning can be utilized as tools to overcome such limitations. This study addresses the challenges that teachers face when drawing three-dimensional geometric shapes in digital environments. Shapes drawn manually in digital environments can often be complex, making it difficult for teachers to create accurate and precise drawings. Deep learning models can assist teachers in correcting drawing errors, thereby providing students with clearer and more comprehensible visuals to facilitate the learning of geometric concepts. The study emphasizes the high accuracy rates achieved using various deep learning models, highlighting their impressive capabilities in accurately classifying geometric shapes.