Şevket Ay
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Şevket Ay has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
- Buket Coşkun
Department of Electrical and Electronics Engineering Yeditepe University Faculty of Engineering İstanbul Türkiye
1 shared publication - Devrim Tarakci
Department of Occupational Therapy, Faculty of Health Sciences, Istanbul Medipol University, Istanbul, Turkey
1 shared publication - Duygun Erol Barkana
Department of Electrical and Electronics Engineering, Yeditepe University, TURKEY
1 shared publication - Ekin Ekinci
Computer Engineering Department, Faculty of Technology, Sakarya University of Applied Sciences, Sakarya, Turkey
1 shared publication - Hilal Bostanci Seçki̇n
Department of Occupational Therapy, Ankara Medipol University Faculty of Health Sciences, Ankara, Türkiye
1 shared publication - İsMail Uzun
INOSENS Bilisim, Gebze Organize Sanayi Bölgesi Technopark, Kocaeli, Türkiye
1 shared publication
Latest Publications
- 3 September 2025
- Published by Kahramanmaras Sutcu Imam University Journal of Engineering Sciences in Kahramanmaraş Sütçü İmam Üniversitesi Mühendislik Bilimleri Dergisi
- Vol. 28 (3) , 1362-1374
A NEW O-NET MODEL FOR AUTOMATED BRAIN TUMOR SEGMENTATION ON MAGNETIC RESONANCE IMAGES
Cancer is a major contributor to global mortality rates, with brain tumors being among the most dangerous types. Early diagnosis and appropriate treatment significantly impact patients' survival rates and quality of life. Moreover, accurately determining the tumor type is crucial for treatment planning and the development of personalized treatment strategies. Therefore, a reliable and effective imaging and analysis method is required to classify brain tumors accurately. In recent years, advancements in deep learning techniques have led to significant progress in medical imaging. This study introduces O-Net, a deep learning model specifically designed to automate brain tumor segmentation. Based on the U-Net architecture, O-Net exhibits a sophisticated structure and depth, enabling superior performance. Experimental evaluations using brain Magnetic Resonance Imaging (MRI) revealed that the O-Net model outperformed other studies in the literature. The model achieved a high accuracy rate with a 90% Dice score. These results demonstrate that O-Net is a reliable tool for the accurate detection and segmentation of brain tumors. The advanced version of the O-Net model can be effectively utilized in future clinical applications for the diagnosis and treatment of brain tumors.
- 3 September 2025
- Published by Kahramanmaras Sutcu Imam University Journal of Engineering Sciences in Kahramanmaraş Sütçü İmam Üniversitesi Mühendislik Bilimleri Dergisi
- Vol. 28 (3) , 1362-1374
A NEW O-NET MODEL FOR AUTOMATED BRAIN TUMOR SEGMENTATION ON MAGNETIC RESONANCE IMAGES
Cancer is a major contributor to global mortality rates, with brain tumors being among the most dangerous types. Early diagnosis and appropriate treatment significantly impact patients' survival rates and quality of life. Moreover, accurately determining the tumor type is crucial for treatment planning and the development of personalized treatment strategies. Therefore, a reliable and effective imaging and analysis method is required to classify brain tumors accurately. In recent years, advancements in deep learning techniques have led to significant progress in medical imaging. This study introduces O-Net, a deep learning model specifically designed to automate brain tumor segmentation. Based on the U-Net architecture, O-Net exhibits a sophisticated structure and depth, enabling superior performance. Experimental evaluations using brain Magnetic Resonance Imaging (MRI) revealed that the O-Net model outperformed other studies in the literature. The model achieved a high accuracy rate with a 90% Dice score. These results demonstrate that O-Net is a reliable tool for the accurate detection and segmentation of brain tumors. The advanced version of the O-Net model can be effectively utilized in future clinical applications for the diagnosis and treatment of brain tumors.
- Open Access
- 21 January 2025
- Published by AVES YAYINCILIK A.Ş. in Archives of Health Science and Research
- Vol. 12 (1) , 1-8
Comparison of Machine Learning Methods to Detect Stress and Participation for Children with Different Special Needs During Serious Game-Based Therapy: A Observational Study
Objective: Serious games have shown promise as therapeutic tools for children with different special needs. However,understanding how children feel and participate in a game is also important. This study used machine learning(ML) methods to classify stress in children with different special needs and their participation in serious game-basedtherapies.Methods: This cross-sectional observational study was conducted at the Pediatric Rehabilitation Laboratory of theDepartment of Occupational Therapy between March and May 2023. Physiological signals such as blood volume pulse,electrodermal activity, and skin temperature were collected from 25 children with obstetric brachial plexus injury, dyslexia, intellectual disabilities, and typically developing children during game therapy. Using these physiological signals,12 ML models were applied to classify children’s stress and participation. Descriptive statistics (mean, SD, frequencies)were used to summarize participant characteristics. Model performance was evaluated using metrics such as accuracy,precision, recall, and F1 score.Results: The results demonstrate that the k-nearest neighbor (KNN) classifier after an autoencoder resulted in the highest F1 scores of 66% and 63% for stress and participation classification, respectively. Furthermore, the eXtreme GradientBoosting (XGBoost) model achieved the highest F1 scores of 91% and 86% for the no-stress and no-participation classifications, respectively. When both minority and majority classes were taken into consideration, using KNN followingan autoencoder yielded better results with average F1 scores of 68% and 65% for stress and no-stress and participationand no participation, respectively.Conclusion: This study shows that ML methods are effective in classifying children’s stress and engagement states usingphysiological signals.Cite this article as: Coskun B, Ay Ş, Erol Barkana D, Uzun İ, Bostancı Seçkin H, Tarakcı D. Comparison of machine learning methods to detect stress and participationfor children with different special needs during serious game-based therapy: A observational study. Arch Health Sci Res. 2025, 12, 0136, doi: 10.5152/ArcHealthSciRes.2025.24136.
- 21 January 2025
- Published by AVES YAYINCILIK A.Ş. in Archives of Health Science and Research
- Vol. 12 (1) , 1-8
Open AccessComparison of Machine Learning Methods to Detect Stress and Participation for Children with Different Special Needs During Serious Game-Based Therapy: A Observational Study
Objective: Serious games have shown promise as therapeutic tools for children with different special needs. However,understanding how children feel and participate in a game is also important. This study used machine learning(ML) methods to classify stress in children with different special needs and their participation in serious game-basedtherapies.Methods: This cross-sectional observational study was conducted at the Pediatric Rehabilitation Laboratory of theDepartment of Occupational Therapy between March and May 2023. Physiological signals such as blood volume pulse,electrodermal activity, and skin temperature were collected from 25 children with obstetric brachial plexus injury, dyslexia, intellectual disabilities, and typically developing children during game therapy. Using these physiological signals,12 ML models were applied to classify children’s stress and participation. Descriptive statistics (mean, SD, frequencies)were used to summarize participant characteristics. Model performance was evaluated using metrics such as accuracy,precision, recall, and F1 score.Results: The results demonstrate that the k-nearest neighbor (KNN) classifier after an autoencoder resulted in the highest F1 scores of 66% and 63% for stress and participation classification, respectively. Furthermore, the eXtreme GradientBoosting (XGBoost) model achieved the highest F1 scores of 91% and 86% for the no-stress and no-participation classifications, respectively. When both minority and majority classes were taken into consideration, using KNN followingan autoencoder yielded better results with average F1 scores of 68% and 65% for stress and no-stress and participationand no participation, respectively.Conclusion: This study shows that ML methods are effective in classifying children’s stress and engagement states usingphysiological signals.Cite this article as: Coskun B, Ay Ş, Erol Barkana D, Uzun İ, Bostancı Seçkin H, Tarakcı D. Comparison of machine learning methods to detect stress and participationfor children with different special needs during serious game-based therapy: A observational study. Arch Health Sci Res. 2025, 12, 0136, doi: 10.5152/ArcHealthSciRes.2025.24136.