Scopus İndeksli Yayınlar Koleksiyonu
Permanent URI for this collectionhttps://hdl.handle.net/20.500.12573/395
Browse
2 results
Search Results
Article Frequency-Based Deep Occlusion Awareness Instance Segmentation(MDPI, 2026-02-26) Guzel, Yasin; Aydin, Zafer; Talu, Muhammed FatihOne major challenge faced by deep learning-based methods that detect target objects in the form of bounding boxes is object occlusion. High degrees of occlusion significantly diminish the accuracy of instance segmentation. Nonetheless, complex-valued Fourier descriptors can robustly represent object boundaries using minimal information. In this study, the impact of integrating Fourier descriptors-renowned for their strong representational capacity-with deep network models (UNet) that exhibit high generalization performance on instance segmentation accuracy was investigated. Within the scope of the research, nine network models were designed based on different strategies for utilizing frequency components. These variants fall into four strategy families: (i) UNet-style spectrum regression on fixed low-frequency windows (FUNet), (ii) magnitude-guided frequency selection/ROI construction (FUNet-Thr, FUNet-BBox), (iii) sequence models over tokenized FFT coefficients (BiLSTM Patch/Sorted), and (iv) encoder-only spectrum predictors with different depth/capacity (EncoderFFT1/2). To fairly evaluate the models' performance in segmenting objects subjected to disruptive factors (e.g., occlusion, blurring, noise), a specialized synthetic dataset was prepared. The task is formulated as single-target (single-instance), single-class segmentation. This dataset, automatically generated according to initial parameter values, contains images of objects moving at various speeds within a single frame. Among these models, the one termed FUNet, which relies on partial matching of central frequency components, achieved the highest segmentation accuracy despite the disruptive effects. Under the challenging Dataset 8 setting, the proposed FUNet achieved the highest overlap-based performance (Dice = 0.9329, IoU = 0.8842) among Attention U-Net, U-Net, and FourierNet, with statistically significant gains confirmed by paired per-image tests.Article A Comparative Study of Unet Variants for Low-Grade Glioma Segmentation in Magnetic Resonance Imaging(Inonu University, 2025-06-25) Guzel, Yasin; Aydin, ZaferBrain tumors originating from glial cells are pathological entities that significantly impact quality of life and are classified based on their malignancy into low-grade gliomas (LGGs) and high-grade gliomas (HGGs). While the more aggressive HGGs have been extensively studied, LGGs are of critical importance for early diagnosis due to their potential progression to HGGs if left untreated. This has driven researchers to develop methods for the rapid and consistent diagnosis of LGGs. In this study, three models—UNet, Transformer UNet, and Super Vision UNet—were comparatively evaluated for the automatic segmentation of LGGs using magnetic resonance imaging (MRI) data. Multimodal MRI scans from 110 patients, retrieved from The Cancer Imaging Archive (TCIA), were used to train the models. Performance was evaluated using Dice Coefficient, Tversky Index, and Intersection over Union (IoU) metrics. The Super Vision UNet achieves the highest Dice (0.9115) and Tversky (0.9154) scores, while the Transformer UNet attains the highest IoU (0.8789). Both advanced models demonstrate superior segmentation performance with lower loss values compared to the conventional UNet. Visual outputs indicate that the modern architectures delineate tumor contours with greater precision. These results highlight the effectiveness and reliability of contemporary UNet-based and Transformer-based architectures in segmenting complex tumor structures such as LGGs. Integrating these models into clinical decision support systems holds promise for enhancing the speed and accuracy of the diagnostic process. © 2025 Elsevier B.V., All rights reserved.