Browsing by Author "Koc, M"

Now showing 1 - 6 of 6
  • No Thumbnail Available
    Item
    Neural network representations for the inter- and intra-class common vector classifiers
    Edizkan, R; Barkana, A; Koc, M; Gulmezoglu, MB; Ashames, MMA; Ergin, S; Fidan, M; Demir, A; Calisir, C; Gerek, ON
    Common Vector Approach (CVA) is a known linear regression-based classifier, which also enables an extension to inter-class discrimination, known as the Discriminative Common Vector Approach (DCVA). The characteristics of linear regression classifiers (LRCs) enable the possibility of a schematic implementation that is similar to the neuron model of artificial neural networks (ANNs). In this work, we explore this schematic similarity to come up with an ANN representation for both CVA and DCVA. The new representation eliminates the need for projection matrices in its implementation, hence significantly reduces the memory requirements and computational complexities of the processes. Furthermore, since the new representation is in a neural style, it is expected to provide a solid and intriguing extension of CVA (and DCVA) by further incorporating adaptation or activation processes to the already successful CVA-based classifiers. (c) 2023 Elsevier Inc. All rights reserved.
  • No Thumbnail Available
    Item
    Are deep learning classification results obtained on CT scans fair and interpretable?
    Ashames, MMA; Demir, A; Gerek, ON; Fidan, M; Gulmezoglu, MB; Ergin, S; Edizkan, R; Koc, M; Barkana, A; Calisir, C
    Following the great success of various deep learning methods in image and object classification, the biomedical image processing society is also overwhelmed with their applications to various automatic diagnosis cases. Unfortunately, most of the deep learning-based classification attempts in the literature solely focus on the aim of extreme accuracy scores, without considering interpretability, or patient-wise separation of training and test data. For example, most lung nodule classification papers using deep learning randomly shuffle data and split it into training, validation, and test sets, causing certain images from the Computed Tomography (CT) scan of a person to be in the training set, while other images of the same person to be in the validation or testing image sets. This can result in reporting misleading accuracy rates and the learning of irrelevant features, ultimately reducing the real-life usability of these models. When the deep neural networks trained on the traditional, unfair data shuffling method are challenged with new patient images, it is observed that the trained models perform poorly. In contrast, deep neural networks trained with strict patient-level separation maintain their accuracy rates even when new patient images are tested. Heat map visualizations of the activations of the deep neural networks trained with strict patient-level separation indicate a higher degree of focus on the relevant nodules. We argue that the research question posed in the title has a positive answer only if the deep neural networks are trained with images of patients that are strictly isolated from the validation and testing patient sets.
  • No Thumbnail Available
    Item
    Histopathological and audiological effects of mechanical trauma associated with the placement of an intracochlear electrode, and the benefit of corticosteroid infusion: prospective animal study
    Malkoc, G; Dalgic, A; Koc, M; Kandogan, T; Korkmaz, S; Ceylan, ME; Inan, S; Olgun, L
    Objective: This study aimed to present the histopathological and audiological effects of mechanical trauma associated with the placement of a model electrode in the scala tympani in rats, and the effects of continuous topical corticosteroid application. Method: The study comprised three groups of rats. The round window membrane was perforated in all three groups and a model electrode was inserted in the round window. Group one received no further treatments. Groups two and three also had an intrathecal microcatheter compatible with a mini-osmotic pump inserted; in group two this was used to release normal saline and in group three the pump released 400 mu g/ml dexamethasone. Results: Dexamethasone infusion given after implantation of the intracochlear model electrode was more effective for preventing hearing loss than the administration of just one dose of dexamethasone. Conclusion: The findings suggest that continuous dexamethasone infusion is beneficial for preventing the loss of hair cells and neurons associated with early and late periods of intracochlear electrode trauma.
  • No Thumbnail Available
    Item
    Determination of bone marrow radiation dose using MIRDOSE3 package program in thyroid cancer patients
    Parlak, Y; Demir, M; Erees, F; Gumuser, F; Uysal, B; Kaya, GC; Koc, M; Ergene, U; Bilgin, E
  • No Thumbnail Available
    Item
    Patterns of care for lung cancer in radiation oncology departments of Turkey
    Demiral, A; Alicikus, ZA; Ugur, V; Karadogan, I; Yoney, A; Andrieu, MN; Yalman, D; Pak, Y; Aksu, G; Ozyigit, G; Ozkan, L; Kilciksiz, S; Koca, S; Caloglu, M; Yavuz, A; Caglar, H; Beyzad-eoglu, M; Igdem, S; Serin, M; Kaplan, B; Koc, M; Korkmaz, E; Celik, OK; Kinay, M
  • No Thumbnail Available
    Item
    Bone Marrow Radiation Dose in Thyroid Carcinoma Patients Treated with Iodine-131
    Parlak, Y; Demir, M; Erees, S; Gumuser, G; Uysal, B; Kaya, GC; Koc, M; Ergene, U; Sayit, E