Yazar "Şengür, Abdulkadir" seçeneğine göre listele
Listeleniyor 1 - 4 / 4
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe Automated Segmentation of Lung Lesions Using Deep Learning: A Study on a Multi-Class CT Dataset(Institute of Electrical and Electronics Engineers Inc., 2025) Kiliç, Murat; Yelman, Abdulkadir; Üzen, Hüseyin; Firat, Hüseyin; Biyikli, Merve; Balikçi Çyçek, Ipek; Şengür, AbdulkadirLung cancer is the leading cause of cancer-related deaths worldwide, making early diagnosis from CT images critically important for patient outcomes. In this study, a unique dataset was collected at Inönü University Turgut Özal Medical Center, consisting of 3526 CT slices from 210 patients manually annotated by expert physicians. The lesions, classified into benign, malignant, and cystic categories, were segmented using seven different deep learning architectures: U-Net, VGG16UNet, EFF-UNet, UNet++, FPNet, PSPNet, and EFF-PSPNet. The models were evaluated based on metrics such as Dice, Jaccard, precision, recall, and FPS. The most successful results were achieved by the EFF-UNet (Dice: 92.14%) and U-Net (Dice: 92.25%) models. While cystic lesions were detected with high accuracy, benign lesions presented the lowest results due to their morphological variability. Visual analyses indicated that some models exhibited uncertainty in class differentiation, and the quality of annotations was found to directly influence model performance. This study demonstrates the potential of deep learning-based segmentation for clinical decision support systems and contributes to the field with its original dataset. © 2025 IEEE.Öğe Grafik Tablet Kullanılarak Makine Öğrenmesi Yardımı ile El Yazısından Cinsiyet Tespiti(2020) Arı, Berna; Arı, Ali; Ucuz, İlknur; Özdemir, Filiz; Şengür, AbdulkadirÖz: Günlük hayatın bir rutini olan el yazısı; ruh hali, kişilik özellikleri ve var olan bazı hastalıklar hakkında ipuçları vermektedir. Bunun yanında adli tıp, tıp ve arkeoloji gibi birçok disiplin el yazısını kendi alanlarına yönelik çalışmalarda sıkça kullanmaktadırlar. Örneğin adli tıp; bazı vakaların aydınlatılmasında el yazısından yaş aralığı ve hangi elin kullanıldığı gibi bilgilere ulaşabilmektedir. Bu çalışmada, el yazısından cinsiyet tespiti yapan bir sistem önerilmiştir. Önerilen sistem el yazısından, el yazısını karakterize edecek bir dizi öznitelik çıkarıp bu öznitelikleri makine öğrenmesi teknikleri kullanarak cinsiyete göre sınıflandırmıştır. El yazılarının kaydedildiği tabletle hem kalemin tablete temas halindeki durumunda hem de harfler ve kelimeler arası geçişte kalemin havada izlediği eğri hareketlerinden öznitelik çıkarılmıştır. Bu öznitelikler sırası ile kalem hızı, ivmesi, yazarken oluşan sarsıntı hareketleri, eğim açısı, yazıdaki kavislenmeler, kalemin havada kalma oranı, kalemin yaptığı basınç değeri ve kalemin yükseklik açısıdır. Sınıflandırıcı olarak da Karar Ağaçları (KA), Naive Bayes (NB), Destek Vektör Makineler (DVM) ve k-en Yakın Komşu (k-EYK) yaklaşımları kullanılmıştır. Deneysel çalışmalarda kullanılan veri setinde toplam 410 örnek mevcut olup, deneysel çalışmaların başarımları doğruluk kriteri ile değerlendirilmiştir. Elde edilen sonuçlara göre en iyi başarımın DVM ile elde edildiği ve doğruluk değerinin de %85,1 olduğu görülmüştür.Öğe MorphMaskFormer: a transformer-based deep segmentation model for multi-class Demirjian stage estimation from panoramic radiographs(Elsevier Inc., 2026) Kıranşal, Melike; Özçelik, Salih Talha Alperen; Aydan, Tuba; Üzen, Hüseyin; Fırat, Hüseyin; Şengür, Abdulkadir; Abdelkarim, Ahmed Z.Objectives This study aims to develop an advanced deep learning model that automatically determines third-molar developmental stages in panoramic radiographs using the Demirjian classification, improving the accuracy and objectivity of dental age estimation for forensic and clinical applications. Study Design A total of 888 panoramic radiographs from individuals aged 7 to 30 were annotated by 2 experts based on Demirjian’s A–H staging system. The proposed model, MorphMaskFormer , is built upon the classical UNet architecture, incorporating a lightweight transformer attention module inspired by Mask2Former. The model performs both binary (tooth/background) and multi-class (A–H stages) segmentation. Its performance was evaluated using IoU, Dice coefficient, Precision, Recall, and inference time, and compared against UNet, ResUNet, DeepLabV3+, PSPNet, and SegNet. Results MorphMaskFormer outperformed all baseline models, achieving a Dice score of 0.9461, IoU of 0.8985, and the fastest inference time at 78.59 ms. In multi-class segmentation, it showed high accuracy for stages A, D, and H, with an overall component accuracy of 72.41%. Conclusions MorphMaskFormer enables precise pixel-level segmentation of dental developmental stages, reducing inter-observer variability and shortening evaluation time. Its high accuracy and efficiency make it a scalable tool that enhances diagnostic confidence and supports critical clinical and forensic age-estimation decisions. © 2026 Elsevier Inc.Öğe Suicide Prediction from Hemogram with Machine Learning(2020) Arı, Berna; Arı, Ali; Şengür, AbdulkadirAbstract: Suicide; It is a phenomenon that we encounter with different frequencies and methods by hosting social, economic and cultural factors at its base. Adolescence, which is an upper step of childhood, contains complex emotions such as hopelessness, loneliness, and depression in its world, and it is a stage in which the risk of suicide is high. It is of great importance to take necessary measures in neutral and imperceptible ways in terms of adolescence and suicide relationship. Blood, which can be easily taken by experts even in a non-severe illness, appears as numerical data with the parametric values that make up its content in laboratories. The hemogram test showing the measurement of blood parameters is used in the diagnosis of many diseases today. In this study, the relationship between the values obtained as a result of the hemogram test and the possibility of suicide of adolescent individuals were investigated. Leukocyte (WBC), erythrocyte (RBC), basophil (BA), eosinophil (EO), lymphocyte (LY), Monocyte (MO), Neutrophil (NE) and Platelet (PLT) of adolescents who have attempted suicide and whose age and gender are known, blood values of mean platelet volume (MPV) and hemoglobin (HGB) levels were evaluated within the designed system. Complete blood count data of 302 individuals who were healthy and suicidal attempts constituting the dataset were pre-processed and the data that would adversely affect the estimated suicide group were removed from the system by considering the references. While making suicide estimation, the high performance bagging trees and the Support Vector Machines separating the members of the two groups with high accuracy were chosen as a result of the joint study of the classification algorithms. It has been shown that by using 260x13 attribute, the classification results can be obtained with BT and Quadratic SVM and 93.5% accurate predictions can be made with BT. Experts will be able to easily find out how high or at which level this probability is, if the individual has any psychological disorders or if the suicide is suspected.
