Sürdürülebilir Kalkınma Amaçları
Yazarlar
- Seker, Dursun Zafer(3)
- Bayram, Bulent(2)
- Catal Reis, Hatice(2)
- Reis, Hatice Catal(2)
- Almasani, Heba(1)
- Avsar, Ozgur(1)
- Bayram, Bülent(1)
- Bozkurt, Salih(1)
- Catal, Hatice(1)
- Corumluoglu, Ozsen(1)
- Demir, Nusret(1)
- Eker, Oktay(1)
- Erdogan, Mustafa(1)
- Fernández-Caballero, A.(1)
- Genemo, Musa(1)
- Ince, Abdulkadir(1)
- Janpaule, Inese(1)
- Kayi, Abdullah(1)
- Papakostas, G.(1)
- Shakya, S.(1)
- Tavares, J.M.R.S.(1)
- Ture, Hayati(1)
- Turk, Veysel(1)
Yayınlar
- BildiriYalnız künye2025Improving Skin Lesion Classification with Pretrained Deep Learning Models
Almasani, Heba; Ture, Hayati
Skin cancer is the most common type of cancer. Cancer can be defined as a disease characterized by the uncontrolled division and spread of cells. As with other types of cancer, early diagnosis and treatment play a crucial role in combating skin cancer. Since diagnosis is made through visual examination, factors such as the level of expertise and image artifacts negatively affect diagnostic accuracy and slow down the process.Computer-aided applications used to analyze such images provide faster and more accurate results. Additionally, these methods eliminate factors that reduce diagnostic accuracy during the classification and segmentation of skin lesions.This study classifies skin cancer to benign or malignant using deep learning and machine learning methods. In particular, Convolutional Neural Networks (CNNs) have shown promising results in medical image classification. The study evaluates six pre-trained CNN models - VGG16, EfficientNetB7, EfficientNetV2L, DenseNet, ConvNeXtBase and MobileNet - for binary skin lesion classification using the ISIC 2019 dataset. The study first preprocesses and balances the data to enhance performance. The results indicate that the ConvNeXtBase model achieved the highest success with accuracy rate of 84.63% after balancing the data that we have.
- MakaleAçık Erişim2025A multi-stage fusion deep learning framework merging local patterns with attention-driven contextual dependencies for cancer detection
Catal Reis, Hatice; Turk, Veysel
Cancer is a severe threat to public health. Early diagnosis of disease is critical, but the lack of experts in this field, the personal assessment process, the clinical workload, and the high level of similarity in disease classes make it difficult. In recent years, deep learning-based artificial intelligence models have shown promise, with the potential to increase diagnosis speed and accuracy. These models attract attention with their automatic learning and adaptation capabilities. In this study, the deep learning-based PADBSRNet model and the PADBSRNet-Vision Transformer (ViT) hybrid method are proposed for the detection of brain tumors and skin and lung cancers. PADBSRNet is a comprehensive deep neural network architecture that integrates separable and traditional convolution layers, multiple attention mechanisms, bidirectional recurrent neural networks, and cross-connections/multi-stage feature fusion strategies. This architecture offers significant advantages in terms of effectively extracting local-global, contextual features and accurately modeling long-term dependencies in image classification tasks. The second proposed approach developed a hybrid method that combines the advantages of the PADBSRNet model and the ViT model. Experimental analysis on medical datasets such as the Figshare Brain Tumor Dataset, IQ-OTH/NCCD Dataset, and Skin Cancer: Malignant vs. Benign Dataset has evaluated the proposed models' performances. According to the experimental results, the PADBSRNet model has shown successful performance with 95.24 %, 99.55 %, and 88.61 % accuracy rates, respectively. The experimental findings show that the proposed deep learning model can effectively learn the complex relationships and hidden patterns of cancer disease, thus producing applicable and effective results in cancer diagnosis.
- BildiriYalnız künye2023Quantum Convolutional Neural Network for Agricultural Mechanization and Plant Disease Detection
Genemo, Musa; Shakya, S.; Tavares, J.M.R.S.; Fernández-Caballero, A.; Papakostas, G.
Agricultural research is essential in addressing the challenges of food production and meeting the needs of a growing population. Severe crop disorders can lead to food insecurity, potentially causing disastrous effects on agricultural production, such as yield losses, reduced crop quality, and even complete crop failures. By leveraging computer vision techniques, it is possible to develop systems that detect signs of crop diseases at an early stage, facilitating prompt intervention and management. Environmental changes, such as rainfall and biosolids, can cause plants to become inflamed by microorganisms. While manual observation and assessment of leaf conditions have traditionally been used for crop disease detection, there is increasing interest in leveraging technology to automate and enhance the process. Manual detection of crop diseases based on leaf conditions is subjective, time consuming, and vulnerable to errors. Utilizing computer vision methods to detect plant diseases helps improve plant health. The concept of applying quantum convolutional neural networks (QCNNs) for agricultural mechanization and plant disease detection is an interesting area of research that combines quantum computing and deep learning techniques. However, it is important to note that quantum computing is still in its early stages, and practical implementations of QCNNs for real-world applications, including agriculture, are not yet widespread. The proposed work of paper focuses on the detection of leaf blight disease by classifying leaves into late blight, early blight, and healthy leaves. This work is divided into two stages: segmentation and classification. For segmentation, MobileNetV2 is used, and an emerging quantum machine learning technique named QCNN is used for classification. The main objective is to provide and propose an effective and fast automated technique to detect and classify blight disease, which will help farmers identify blight disease easily. The proposed model achieved a classification accuracy of 0.969. Furthermore, the results were further validated against recent findings in the field. The presented outcome implies that the model’s performance is remarkable.
- MakaleYalnız künye2018An Extended Approach of Particle Swarm Optimization for Shoreline Extraction from RASAT Imagery
Catal Reis, Hatice; Bayram, Bülent; Bozkurt, Salih; Ince, Abdulkadir; Demir, Nusret; Seker, Dursun Zafer
This study presents an extended shoreline detection approach from pansharpened images of Turkish RASAT satellite which covers red, green and blue bands of the optical spectrum with 15 m ground resolution and panchromatic band with 7.5 m spatial resolution. The Lake Ercek of Turkey has been selected as the study area, which is a tectonic lake and home to a variety of water birds. The satellite images of the lake taken in 2013 and 2014 were considered for analysis. The proposed shoreline extraction system consists of a sequence of image processing steps in which simple linear clustering (SLIC) and particle swarm optimization (PSO) are the main components. SLIC was used to create superpixels that form basis for object-based image analysis while PSO was employed for classifying objects into corresponding classes. The resulting images still contained unwanted artefacts; therefore, a post-processing step was performed to improve the accuracy of segmentation by applying thresholding, morphological processing, and manual editing for noise removal. The proposed framework was applied on two temporal RASAT images to test the variations of defined parameter settings. The success of the proposed system was to obtain shorelines with satisfying accuracy without using NIR band. Finally, the extracted shorelines were vectorised and compared with manually digitized shorelines from pansharpened satellite images for accuracy assessment.
- BildiriYalnız künye2017THE ROLE OF NATIONAL AND INTERNATIONAL GEOSPATIAL DATA SOURCES IN COASTAL ZONE MANAGEMENT
Bayram, Bulent; Avsar, Ozgur; Seker, Dursun Zafer; Kayi, Abdullah; Erdogan, Mustafa; Eker, Oktay; Janpaule, Inese; Reis, Hatice Catal
Coastline changes are increasing rapidly due to both natural and human effects. The expansion of touristic, industrial and culture fishing establishments through coastal areas has brought about uncontrolled and unplanned urbanization. In this study, the coastline in Karasu district of Turkey has been extracted by using Unmanned Aerial Vehicle (UAV) images. For this purpose; 7 cm ground sample distance (GSD) UAV images taken in 2013 by Gatewing-X100 by 40% side and 70% forward overlap was used as post images. Produced ortho images were used as the base map for extraction of the coastline. An object-oriented approach has been applied to capture the shoreline from these ortho images. In the presented study, the eCognition object-oriented fuzzy image processing software has been used. This commercial software has the ability to develop custom tools for image classification in addition to its standard object feature tools. Customized arithmetic features have been used to achieve more accurate shoreline segmentation results. Three main object classes have been created as sea, shoreline buffer and land to extract shoreline. The segmentation results were converted into dxf vector data format. The results were compared with manual digitizing of 55 blind readers. An algorithm has been developed by using Matlab to analyze the differences between object-oriented classification and manual digitizing results. Results were evaluated according to students' gender, age, spent time, used software and courses taken. The root mean square error (RMSE) was calculated as 7.15 m.
- MakaleYalnız künye2016A semiautomatic segmentation approach to biometric measurement of the talus bone of sedentary women and ballerinas using CT images
Reis, Hatice Catal; Bayram, Bulent; Seker, Dursun Zafer
Background: Ballet produces much stress on bones in the feet of ballerinas. Monitoring and detecting talus bone deformation is important for their quality of health and profession. Objectives: To determine differences in the talus bone between ballerinas and sedentary women. Methods: We evaluated biometric differences in the talus bone of 5 ballerinas referenced to 5 similar sedentary women recruited into the present study. We acquired 20 multidetector computed tomographic images including right and left feet. Semiautomatic region-based image processing using 3D-Doctor (Able Software Corp) was used to create three-dimensional (3D) virtual models of the talus bones. Biometric measurements were made on the 3D models and statistical analysis conducted. Results: The mean talus bone length of ballerinas was 3.37 cm (SD 0.12; range 3.11 to 3.52). The talus bone length of sedentary women was 3.29 cm (SD 0.16; range 3.04 to 3.65), and tended to be shorter than that of the ballerinas (P = 0.08, t test). However, the test was insufficiently powered. The mean volume of the talus bones from the ballerinas was 24.8 cm3 (SD 0.83) and smaller than that from the sedentary women, 26.9 cm3 (SD 1.25) at P < 0.001 (t = 4.38, 18 degrees of freedom. Difference 2.1. Two-tailed 95% confidence interval for difference of means: 1.08 to 3.08). We found less variation in the feet of ballerinas than sedentary women. Conclusions: Volumetric measurements show that feet of ballerinas are smaller and retain similar shape and size than the irregular feet of sedentary women.
- BildiriYalnız künye2012Three-dimensional modeling of the sesamoid bones
Catal, Hatice; Corumluoglu, Ozsen
In this research, digital image processing techniques was used to create 3-dimensional (3D) virtual model of sesamoid bones. As study materials, Ballerina's and sedentary women's axial MDCT (Multidetector Computed Tomography) feet images were utilized. For the determination bone boundaries in each slides, image preprocessing and then segmentation steps were followed. All these processing steps was done in 3DDoctor software. All the measurements were obtained over the 3D model formed. Various statistical ratios were computed regarding the area, volume and length measurements. The comparison done in the study was performed by using these ratios for sedentary and ballerinas. As an outcome, some differences were found in sesamoid bones of ballerina and sedentary women.