Yazar "Karaoglu, Ahmet" seçeneğine göre listele

Listeleniyor 1 - 3 / 3
  • [ X ]
    Öğe
    A New Approach for Improving Biodiesel Conversion Efficiency: A Stacking Ensemble Model Based on Linear Regression Approach with GAN-Enhanced
    (Springer Heidelberg, 2025) Karaoglu, Ahmet; Soyler, Huseyin
    This study employs a Linear Regression-based stacking ensemble learning approach as a novel method to enhance biodiesel conversion efficiency. Initially, a dataset derived from the literature was used to train an ensemble model that combines predictions from Random Forest, XGBoost, and Deep Neural Network (DNN) through a Linear Regression-based fusion approach. This model outperformed individual models (Random Forest: - 0.16, XGBoost: - 0.67, and DNN: 0.36) by achieving an R2 score of 0.45. To further improve model performance, 4900 synthetic data samples were generated and integrated into the dataset. Leveraging the stacking ensemble learning approach with this expanded dataset, the model demonstrated a significant improvement in predictive accuracy, achieving an R2 score of 0.81. This corresponds to an approximate 4% increase in performance compared to individual models (Random Forest: 0.78, XGBoost: 0.78, and DNN: 0.77), highlighting the effectiveness of ensemble learning in optimizing biodiesel conversion efficiency. Additionally, the model exhibited high accuracy with low error rates (MAE: 1.16 and MAPE: 1.24%), effectively compensating for the weaknesses of individual models and providing more stable and generalized predictions. To the best of our knowledge, this is the first study to incorporate a Linear Regression-based stacking method to enhance biodiesel conversion efficiency. These findings underscore the potential of ensemble learning techniques and synthetic data integration in improving renewable fuel efficiency. Future research can further enhance model performance by incorporating larger datasets and exploring more advanced ensemble strategies.
  • [ X ]
    Öğe
    An enhanced deep learning model for detection and classification of dental caries in panoramic radiographs
    (Springer Science and Business Media Deutschland GmbH, 2026) Ozdemir, Dilara; Ozcan, Caner; Karaoglu, Ahmet; Pekince, Adem; Yasa, Yasin; Kazangirler, Buse Yaren; Meseci, Elif
    Early diagnosis of dental caries has become increasingly important in recent years. It reduces irreversible tooth loss, treatment costs and treatment time. However, since the examination of dental caries is carried out visually by experts on radiographic images, the analysis process is quite exhausting for the experts. In addition, visual analysis may miss early-stage caries due to the workload in the clinical environment. In this study, an automatic caries diagnosis system is proposed to support the expert and to reduce the clinical workload by using panoramic images. The proposed DenseNet121-C model, based on deep learning models, generates results with its configured classifier for caries detection. The dataset prepared for the study includes 14498 tooth images automatically cropped from panoramic images. The proposed model achieved the highest performance on the test set with 93.17% accuracy, 89.43% precision, 85.84% recall, and 87.49% F1-score. Considering the high results of the current study, dentists can spend more time on treatment during dental examinations, thanks to the model’s ability to distinguish between caries and non-caries teeth. The results obtained were compared with the Mask R-CNN results. In addition, the performance of the deep learning architectures was investigated on an unbalanced dataset. © The Author(s) 2026.
  • [ X ]
    Öğe
    Numbering teeth in panoramic images: A novel method based on deep learning and heuristic algorithm
    (Elsevier - Division Reed Elsevier India Pvt Ltd, 2023) Karaoglu, Ahmet; Ozcan, Caner; Pekince, Adem; Yasa, Yasin
    Dental problems are one of the most common health problems for people. To detect and analyze these problems, dentists often use panoramic radiographs that show the entire mouth and have low radiation exposure and exposure time. Analyzing these radiographs is a lengthy and tedious process. Recent studies have ensured dental radiologists can perform the analyses faster with various artificial intelligence sup-ports. In this study, the numbering performance of Mask R-CNN and our heuristic algorithm-based method was verified on panoramic dental radiographs according to the Federation Dentaire Internationale (FDI) system. Ground-truth labelling of images required for training the deep learning algorithm was performed by two dental radiologists using the web-based labelling software DentiAssist created by the first author. The dataset was created from 2702 anonymized panoramic radio-graphs. The dataset is divided into 1747, 484, and 471 images, which serve as training, validation, and test sets. The dataset was validated using the k-fold cross-validation method (k = 5). A three-step heuristic algorithm was developed to improve the Mask R-CNN segmentation and numbering results. As far as we know, our study is the first in the literature to use a heuristic method in addition to traditional deep learning algorithms in detection, segmentation and numbering studies in panoramic radiography. The experimental results show that the mAp (@IOU = 0.5), precision, recall and f1 scores are 92.49%, 96.08%, 95.65% and 95.87%, respectively. The results of the learning-based algorithm were improved by more than 4%. In our research, we discovered that heuristic algorithms could improve the accuracy of deep learning-based algorithms. Our research will significantly reduce dental radiologists' workload, speed up diagnostic processes, and improve the accuracy of deep learning systems.(c) 2022 Karabuk University. Publishing services by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).