Yazar "Hanbay, Kazim" seçeneğine göre listele
Listeleniyor 1 - 18 / 18
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe Channel-boosted multi-scale R-CNN for accurate and real-time ship and land detection in complex SAR scenes(Elsevier Sci Ltd, 2026) Hanbay, Kazim; Ozcelik, Salih Taha Alperen; Altin, Mustafa; Uzen, HuseyinAccurate ship detection in Synthetic Aperture Radar (SAR) imagery is crucial for maritime surveillance but remains challenging due to small target sizes, speckle noise, and complex sea-surface backgrounds. While most existing methods focus exclusively on identifying ships, our approach also achieves reliable detection of land areas, providing an additional contribution to the literature. This study introduces CBM-RCNN (Channel-Boosted Multi-Scale R-CNN), a novel deep learning architecture that integrates Convolutional Block Attention Modules (CBAM) and a Bidirectional Feature Pyramid Network (BiFPN) on a ResNet50 backbone. CBAM enhances both spatial and channel-level feature representation, enabling reliable detection of small vessels, while BiFPN fuses multi-scale features bidirectionally, improving accuracy across vessels of different sizes and positions. CBM-RCNN was evaluated against standard Faster R-CNN and YOLOv8 models across diverse maritime scenes, including simple, densely populated, and visually complex scenarios. The model demonstrated superior detection accuracy, balanced class-specific performance, and strong generalization. It effectively resolves overlapping vessels, distinguishes ships from coastal structures, and maintains robustness under challenging SAR-specific noise conditions. Importantly, it achieves inference speeds suitable for near-real-time applications, highlighting practical applicability. By combining attention-driven refinement with multi-scale feature aggregation, CBM-RCNN addresses limitations of prior methods, particularly in small object recognition, complex scene generalization, and simultaneous land detection. This architecture provides a robust framework for automated maritime monitoring and offers a foundation for future improvements in large-scale SAR-based ship detection and environmental surveillance.Öğe CNN-Based Fabric Defect Detection System on Loom Fabric Inspection(E.U. Printing And Publishing House, 2022) Talu, Muhammed Fatih; Hanbay, Kazim; Varjovi, Mahdi HatamiFabric defect detection is generally performed based on human visual inspection. This method is not effective and it has various difficulties such as eye delusion and labor cost. To deal with these problems, machine learning and computer vision-based intelligent systems have been developed. In this paper, a novel real-time fabric defect detection system is proposed. The proposed industrial vision system has been operated in real-time on a loom. Firstly, two fabric databases are constructed using real fabric images and new defective patch capture (DPC) algorithm. One of the main objectives in this study is to develop a CNN architecture that focuses only on fabric defect detection. One of the most unique aspects of the study is to detect defective pixel regions of fabric images with Fourier analysis on a patch-based and integrate it with deep learning Thanks to the novel developed fast Fourier transform-based DPC algorithm, defective texture areas become visible and defect-free areas are suppressed, even on complex denim fabric textures. Secondly, an appropriate convolution neural networks (CNN) model is developed. Thus the new dataset dataset is refined using negative mining method and CNN model. However, traditional feature extraction and classification approaches are also used to compare classification performances of deep models and traditional models. Experimental results show that our proposed CNN model integrated with negative mining can classify the defected images with high accuracy. Also, the proposed CNN model has been tested in real-time on a loom, and it achieves 96.5% detection accuracy. The proposed model obtains better accuracy and speed performance in terms of detection accuracy with a much smaller model size.Öğe Continuous rotation invariant features for gradient-based texture classification(Academic Press Inc Elsevier Science, 2015) Hanbay, Kazim; Alpaslan, Nuh; Talu, Muhammed Fatih; Hanbay, Davut; Karci, Ali; Kocamaz, Adnan FatihExtracting rotation invariant features is a valuable technique for the effective classification of rotation invariant texture. The Histograms of Oriented Gradients (HOG) algorithm has been proved to be theoretically simple, and has been applied in many areas. Also, the co-occurrence HOG (CoHOG) algorithm provides a unified description including both statistical and differential properties of a texture patch. However, HOG and CoHOG have some shortcomings: they discard some important texture information and are not invariant to rotation. In this paper, based on the original HOG and CoHOG algorithms, four novel feature extraction methods are proposed. The first method uses Gaussian derivative filters named GDF-HOG. The second and the third methods use eigenvalues of the Hessian matrix named Eig(Hess)-HOG and Eig(Hess)-CoHOG, respectively. The fourth method exploits the Gaussian and means curvatures to calculate curvatures of the image surface named GM-CoHOG. We have empirically shown that the proposed novel extended HOG and CoHOG methods provide useful information for rotation invariance. The classification results are compared with original HOG and CoHOG algorithms methods on the CUReT, KTH-TIPS, KTH-TIPS2-a and UIUC datasets show that proposed four methods achieve best classification result on all datasets. In addition, we make a comparison with several well-known descriptors. The experiments of rotation invariant analysis are carried out on the Brodatz dataset, and promising results are obtained from those experiments. (C) 2014 Elsevier Inc. All rights reserved.Öğe Fabric Defect Detection Methods for Circular Knitting Machines(Ieee, 2015) Hanbay, Kazim; Talu, Muhammed Fatih; Ozguven, Omer Faruk; Ozturk, DursunIn this paper, an online fabric defect detection system that can detect fabric defects which may occur during the fabric product in knitting machines is introduced. This system mainly includes three steps: 1) Construction of a defected/defect-free fabric database; 2) Obtaining and classification of the feature vectors; 3) Online working on embedded system. This study only contains information about the first two stages. In the first stage, 3242 'defected' and '5923' defect-free images were acquired by using a conveyor system which has line scan camera and linear light. In the second stage, filtering, feature extraction (wavelet transform, co-occurrence matrix and CoHOG) and classification (YSA) processes were carried out. As a result, obtaining the feature vectors through wavelet transform has reduced computation cost by 53% and also has successfully provided the classification of the defects by 90%.Öğe Fabric defect detection systems and methods-A systematic literature review(Elsevier Gmbh, 2016) Hanbay, Kazim; Talu, Muhammed Fatih; Ozguven, Omer FarukThis paper presents a comprehensive literature review of fabric defect detection methods First, it briefly explains basic image acquisition system components such as camera and lens. Defect detection methods are categorized into seven classes as structural, statistical, spectral, model-based, learning, hybrid and comparison studies. These methods are evaluated according to such criteria as the accuracy, the computational cost, reliability, rotating/scaling invariant, online/offline ability to operate and noise sensitivity. Strengths and weaknesses of each approach are comparatively highlighted. In addition, the availability of utilizing methods for weaving and knitting in machines is investigated. The available review studies do not provide sufficient information about fabric defect detection systems for readers engaged in research in the area of textile and computer vision. A set of examination for efficient establishment of image acquisition system are added. In particular, lens and light source selection are mathematically expressed. (C) 2016 Elsevier GmbH. All rights reserved.Öğe Fractional-order gradient based local binary pattern for texture classification(Pergamon-Elsevier Science Ltd, 2025) Alpaslan, Nuh; Hanbay, KazimThe local binary patterns method plays an efficient role in texture classification and feature extraction. These approaches extract textural features by using the neighboring pixel values. The single or joint histogram of the texture image is constructed from the LBP features obtained from local relationships. In this study, a method of utilizing fractional derivative information effectively has been proposed for classifying color texture images. The magnitude of the fractional horizontal and vertical derivatives obtained with Gaussian derivative filters are integrated into the ACS-LBP method. The magnitude information of the fractional derivatives of local texture patterns has been modeled according to the relationship between neighboring pixels. The computed derivative information has been incorporated into the ACS-LBP model to effectively encode the local pixel relationship. In order to maintain, these fractional-order edge and texture transition detection operators provide both high robustness and continue to detect small textural details. To accomplish these capabilities, the fractional-order parameter is tuned to target particular pixel transition frequencies. This gives the proposed LBP method greater latitude in selecting the fractional-order mask. An additional degree of freedom in designing various masks is provided by the fractional-order parameter. The developed model has been evaluated on widely used texture databases. It also has been compared with existing LBP and deep learning models in terms of different performance metrics. The proposed method has shown significant advantages over up to date methods in both classification accuracy and execution time.Öğe A new standard error based artificial bee colony algorithm and its applications in feature selection(Elsevier, 2022) Hanbay, KazimFeature selection is a basic task for pattern recognition and classification. It enhances the performance of the classification algorithms with the help of removing the redundant features. Thanks to eliminating irrelevant features, the computational time is decreased. Thus, intensive works have been carried out in this area. This paper proposes a new standard error-based artificial bee colony (SEABC) algorithm for the feature selection problem, which is developed by integrating standard error-based new solution search mechanisms into the original artificial bee colony algorithm. The SEABC algorithm is used for feature selection. Shannon entropy function is used to serve as the objective function of the SEABC algorithm. Thirteen datasets are used from UCI machine learning datasets. Features are selected according to Shannon conditional entropy values and then a threshold process is implemented to find their best relevant subset. Support Vector Machines (SVMs) and k-Nearest Neighbor (KNN) are used as the optimal classifiers. The proposed SEABC algorithm is compared with genetic algorithm (GA), particle swarm optimization (PSO), ABC, improved ABC (I-ABC), Gbest-guided ABC (GABC), and PS-ABC algorithms. In general, it is observed that the SEABC algorithm achieves better classification results than other wellknown algorithms.(c) 2021 The Author. Published by Elsevier B.V. on behalf of King Saud University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).Öğe Noise Reduced SAR Ship Database(Institute of Electrical and Electronics Engineers Inc., 2024) Hanbay, Kazim; Uzen, Huseyin; Ozdemir, Taha Burak; Ercelik, CetinSynthetic aperture radar (SAR) images are used extensively in agricultural applications, coastal boundary detection and object recognition. This imaging technology provides desired results in many challenging applications due to its ability to provide images with appropriate resolution in harsh weather and climate conditions. In this study, an image database was created to detect ships from SAR images. The images were preprocessed in accordance with the literature and made suitable for ship detection methods. The noise in the images was reduced with a deep learning-based architecture. Using this database, image processing and machine learning methods were used to develop ship detection methods. © 2024 IEEE.Öğe A novel active contour model for medical images via the Hessian matrix and eigenvalues(Pergamon-Elsevier Science Ltd, 2018) Hanbay, Kazim; Talu, Muhammed FatihThis paper presents a new level set formulation for active contour models (ACM). We propose the idea of integrating the eigenvalue information of Hessian matrix into the level set function. By this new level set function, the principal curvature information of images is used to enhance the ability of segmenting boundary regions. The advantages of our model are as follows: firstly, the interior and exterior object boundaries can be segmented with the initial contour being anywhere in the input image. Secondly, this method can work with heterogeneous images. Thirdly, the proposed model can produce smooth and right boundaries of objects having vital importance in medical operations. Extensive experiments demonstrate that the proposed model can obtain better segmentation results. (C) 2018 Elsevier Ltd. All rights reserved.Öğe A NOVEL TEXTURE CLASSIFICATION METHOD BASED ON HESSIAN MATRIX AND PRINCIPAL CURVATURES(Ieee, 2014) Alpaslan, Nuh; Hanbay, Kazim; Hanbay, Davut; Talu, M. FatihIn this study, in order to obtain similar effect with conventional gradient operation and extract more robust feature for texture, we use the principal curvature informations instead of the gradient calculation. Through this methods, sharp and important informations about the texture images were obtained by analyzing images of the second order. Considering the classification results obtained, it is shown that the proposed method improve the performance of original CoHOG and HOG feature extraction methods. As a result of experiments on datasets with different characteristics, it is seen that, the proposed method has higher classification performance.Öğe Panoramic Dental Segmentation with a Novel Approach Based on MONAI U-Net and Sliding Window Inference(Institute of Electrical and Electronics Engineers Inc., 2025) Erçelik, Çetin; Hanbay, KazimTooth segmentation plays a vital role in dental diagnosis and treatment planning. Accurate delineation of dental structures from imaging modalities such as panoramic radiographs forms the foundation for various clinical applications, including orthodontic assessment, implant planning, and disease diagnosis. However, this task is considered complex due to challenges such as image noise, low contrast, overlapping anatomical structures, and missing teeth. As a result, traditional image processing techniques often fall short, giving rise to the prominence of deep learning-based segmentation methods. In this study, four distinct deep learning architectures were evaluated to enhance the performance of dental segmentation in panoramic radiographs: SegNet, U-Net, MONAI U-Net, and the proposed MONAI U-Net with Sliding Window Inference (SWI). Notably, while SWI is commonly employed during validation and testing stages in the literature, it was integrated into the training phase in this study. This modification enabled the model to process smaller patches instead of the full image, thereby preserving fine-grained details and improving segmentation accuracy. Experimental results demonstrated that the proposed MONAI U-Net + SWI model outperformed the other models in both Intersection over Union (IoU) and Dice coefficient metrics. These findings suggest that the proposed approach offers a reliable solution for clinically sensitive tasks such as dental segmentation. © 2025 IEEE.Öğe Principal curvatures based rotation invariant algorithms for efficient texture classification(Elsevier, 2016) Hanbay, Kazim; Alpaslan, Nuh; Talu, Muhammed Fatih; Hanbay, DavutThe histograms of oriented gradients (HOG) and co-occurrence HOG (CoHOG) algorithms are simple and intuitive descriptors. However, the HOG and CoHOG algorithms based on gradient computation still have some shortcomings: they ignore meaningful textural properties and are unstable to noise. In this paper, two new efficient HOG and CoHOG methods are proposed. The proposed algorithms are based on the Gaussian derivative filters, and the feature vectors are obtained by means of principal curvatures. The feature vectors are rotation invariant by means of the rotation invariance characteristic of principal curvatures (i.e. eigenvalues). The experimental results on the CUReT, ICTH-TIPS, KTH-11PS2-a, UIUC, Brodatz album, Kylberg and Xu datasets confirm that the developed algorithms have higher classification rates than state-of-the-art texture classification methods. The classification results also demonstrate that the developed algorithms are more stable to noise and rotation than the original HOG and CoHOG algorithms. (C) 2016 Elsevier B.V. All rights reserved.Öğe Real time fabric defect detection by using fourier transform(Gazi Univ, Fac Engineering Architecture, 2017) Hanbay, Kazim; Talu, Muhammed Fatih; Ozguven, Omer FarukIn fabric productions, real time defect detection systems are needed to detect the surface defects. There is currently no real time defect detection system in knitting fabric production. In this paper, a real time fabric defect detection system is developed and is tested on circular knitting machine. Textural features of fabric image are extracted based on Fourier transform. These textural features are seven and are calculated from the horizontal and vertical directions of Fourier frequency spectrum of the fabric image. The performance of the proposed method is firstly evaluated off-line through experiments based on comprehensive fabric database. The proposed method obtains superior performance, which also proves its utility in real-time inspection. Secondly, a real time machine-vision system has been designed for an efficient detection of the fabric defects under industrial conditions. Real time defect detection system is tested automatically by analyzing fabric images captured by a line scan camera. Experimental results show that the proposed detection model can successfully detect common circular knitting fabric defects.Öğe Real time fabric defect detection system on Matlab and C plus plus /Opencv platforms(Ieee, 2017) Hanbay, Kazim; Golgiyaz, Sedat; Talu, Muhammed FatihIn industrial fabric productions, real time systems are needed to detect the fabric defects. This paper presents a real time defect detection approach which compares the time performances of Matlab and C++ programming languages. In the proposed method, important texture features of the fabric images are extracted using CoHOG method. Artificial neural network is used to classify the fabric defects. The developed method has been applied to detect the knitting fabric defects on a circular knitting machine. An overall defect detection success rate of 93% is achieved for the Matlab and C++ applications. To give an idea to the researches in defect detection area, real time operation speeds of Matlab and C++ codes have been examined. Especially, the number of images that can be processed in one second has been determined. While the Matlab based coding can process 3 images in 1 second, C++/Opencv based coding can process 55 images in 1 second. Previous works have rarely included the practical comparative evaluations of software environments. Therefore, we believe that the results of our industrial experiments will be a valuable resource for future works in this area.Öğe Real-Time Detection of Knitting Fabric Defects Using Shearlet Transform(Ege Univ, 2019) Hanbay, Kazim; Talu, Muhammed Fatih; Ozguven, Omer Faruk; Ozturk, DursunThis paper proposes a vision-based fabric inspection system for the circular knitting machine. Firstly, a comprehensive fabric database called Fabric Defect Detection Database (FDDD) are constructed. To extract significant features of fabric images, shearlet transform is used. Means and variances are calculated from all subbands and combined into a high-dimensional feature vector. The proposed system is evaluated on a circular knitting machine in a textile factory. The real-time performance analysis is only carried out by inspecting single jersey knitted fabric. Our proposed system achieves the highest accuracy of 94.0% in the detection of single jersey knitting fabric defects.Öğe SAR Ship Detection Based on Gaussian Probability and Eigenvalue Analysis(Ieee-Inst Electrical Electronics Engineers Inc, 2025) Hanbay, KazimSynthetic Aperture Radar (SAR) images are frequently used because they provide optimal image quality in all weather conditions. Nevertheless, SAR ship detection has two difficulties. One is coherent speckle noise, which raises false alarms and confuses ships with similar objects. This letter proposes an efficient ship detector for low contrast, inshore and dense targets. First, to accurately eliminate the land areas and speckle noise, the hessian matrix and eigenvalues of the images were calculated. The largest eigenvalue information was given as input to the Gaussian function and the standard deviation and average images of the images were calculated. Then, the standard deviation and average images were combined with a probabilistic approach to obtain an image that highlights the ship regions. Morphological operations and connected component analysis were performed on this image. Experimental results showed that the proposed method provides both accurate and faster results.Öğe Segmentation of SAR images using improved artificial bee colony algorithm and neutrosophic set(Elsevier Science Bv, 2014) Hanbay, Kazim; Talu, M. FatihThis paper proposes a novel synthetic aperture radar (SAR) image segmentation algorithm based on the neutrosophic set (NS) and improved artificial bee colony (I-ABC) algorithm. In this algorithm, threshold value estimation is considered as a search procedure that searches for a proper value in a grayscale interval. Therefore, I-ABC optimization algorithm is presented to search for the optimal threshold value. In order to get an efficient and powerful fitness function for I-ABC algorithm, the input SAR image is transformed into the NS domain. Then, a neutrosophic T and I subset images are obtained. A co-occurrence matrix based on the neutrosophic T and I subset images is constructed, and two-dimensional gray entropy function is described to serve as the fitness function of I-ABC algorithm. Finally, the optimal threshold value is quickly explored by the employed, onlookers and scouts bees in I-ABC algorithm. This paper contributes to SAR image segmentation in two aspects: (1) a hybrid model, having two different feature extraction methods, is proposed. (2) An optimal threshold value is automatically selected by maximizing the separability of the classes in gray level image by incorporating a simple and fast search strategy. The effectiveness of the proposed algorithm is demonstrated by application to real SAR images. (C) 2014 Elsevier B.V. All rights reserved.Öğe Ship Detection in SAR Images Using CLAHE-Enhanced YOLOv8x(Institute of Electrical and Electronics Engineers Inc., 2025) Erçelik, Çetin; Hanbay, Kazim; Özbek, Hasan CanDetection of ships in Synthetic Aperture Radar (SAR) imagery remains a challenging research problem due to complex background structures, low contrast, and speckle noise. In this study, an effective object detection system is proposed by integrating CLAHE (Contrast Limited Adaptive Histogram Equalization)-based preprocessing with the YOLOv8x deep learning architecture to accurately detect both small and large ships in SAR images. The CLAHE algorithm enhances local contrast and reduces noise and distortions in the images, enabling the network to learn more effectively. Subsequently, the YOLOv8x model is trained on these enhanced images. The experimental evaluation on a custom dataset derived from high-resolution Sentinel-1 SAR imagery reveals that the proposed model outperforms both YOLOv8x and YOLOv8n, achieving 62.42% mAP 50,73.83% precision, and 49.66% recall. Visual outputs further confirm that the proposed model significantly improves the detection of small ships. These findings indicate that the CLAHE+YOLOv8x architecture offers an effective and practical approach for ship detection in SAR imagery. © 2025 IEEE.
