Yazar "Oksuz, Cosku" seçeneğine göre listele

Listeleniyor 1 - 3 / 3
  • Küçük Resim
    Öğe
    An integrated convolutional neural network with attention guidance for improved performance of medical image classification
    (Springer London Ltd, 2024) Oksuz, Cosku; Urhan, Oguzhan; Gullu, Mehmet Kemal
    Today, it becomes essential to develop computer vision algorithms that are both highly effective and cost-effective for supporting physicians' decisions. Convolutional Neural Network (CNN) is a deep learning architecture that enables learning relevant imaging features by simultaneously optimizing feature extraction and classification phases and has a high potential to meet this need. On the other hand, the lack of low- and high-level local details in a CNN is an issue that can reduce the task performance and prevent the network from focusing on the region of interest. To tackle this issue, we propose an attention-guided CNN architecture, which combines three lightweight encoders (the ensembled encoder) at the feature level to consolidate the feature maps with local details in this study. The proposed model is validated on the publicly available data sets for two commonly studied classification tasks, i.e., the brain tumor and COVID-19 disease classification. Performance improvements of 2.21% and 1.32%, respectively, achieved for brain tumor and COVID-19 classification tasks confirm our assumption that combining encoders recovers local details missed in a deeper encoder. In addition, the attention mechanism used after the ensembled encoder further improves performance by 2.29% for the brain tumor and 6.13% for the COVID-19 classification tasks. Besides that, our ensembled encoder with the attention mechanism enhances the focus on the region of interest by 4.4% in terms of the IoU score. Competitive performance scores accomplished for each classification task against state-of-the-art methods indicate that the proposed model can be an effective tool for medical image classification.
  • Küçük Resim Yok
    Öğe
    Intelligent Diagnosis and Treatment Systems
    (CRC Press, 2024) Oksuz, Cosku; Yurdem, Betul; Gullu, Mehmet Kemal
    After the first unknown case of pneumonia emerged in China in December 2019, cases reported worldwide soon increased. The new type of coronavirus called SARS-CoV-2, which was determined to be the source of unknown pneumonia, caused the situation to be declared a pandemic within four months. After the past two years, the pandemic continued with the new mutations of the virus. The protracted pandemic has drastically impacted the whole world in many ways. The RT-PCR, which is accepted as the standard testing, has been used for detecting and isolating patients. Especially the high rates of false negatives for the RT-PCR test caused the need to develop alternative tools that are extremely sensitive. Therefore, many methods have been developed adopting machine- and deep-learning-based methods for recognizing COVID-19 disease over medical images. Many of these proposed intelligent systems are based on image-processing methods. More specifically, the researchers are rivaling in a manner to design deep learning based image-processing architectures for capturing the disease patterns effectively. In the scope of this study, the intelligent diagnosis methods proposed in the literature specifically for COVID-19 detection are overviewed by giving the logic behind and conceptualizing them. © 2025 Mustafa Berktas, Abdulkadir Hiziroglu, Ahmet Emin Erbaycu, Orhan Er and Sezer Bozkus Kahyaoglu.
  • Küçük Resim Yok
    Öğe
    A Lightweight Deep Model for Brain Tumor Segmentation
    (Ieee, 2021) Oksuz, Cosku; Urhan, Oguzhan; Gullu, Mehmet Kemal
    Brain tumors are one of the major causes of increasing deaths worldwide. It is important to correctly identify cancerous tissues by experts in order to make correct treatment planning and to increase patient survival rates. However, manually tracking and segmentation of cancerous tissues in many sections of volumetric MR data is an error-prone and time-consuming process. Developments in the field of deep learning in recent years allow the tasks performed by humans to be performed with higher accuracy and speeds through the developed automatic systems. In this study, a deep learning-based light-weighted model with 6.78M parameters is proposed for the classification of cancerous tissues in the brain. Cross-validation of the proposed method on a public data set results in 84.61%, 82.54%, and 87.15% Boundary F-1, mean IoU, and mean accuracy, respectively, shows the robustness of the proposed model.