Fatih Nar

Ankara Yildirim Beyazit University | AYBU · Department of Computer Engineering

Auto-cropping, the process of automatically adjusting the boundaries of an image to focus on the region of interest, is crucial to improving the diagnostic quality of dental panoramic radiographs. Its importance lies in its ability to standardize the size of different input images with minimal loss of information, thus ensuring consistency and impr...

Throughout the evolution of machine learning, the size of models has steadily increased as researchers strive for higher accuracy by adding more layers. This escalation in model complexity necessitates enhanced hardware capabilities. Today, state-of-the-art machine learning models have become so large that effectively training them requires substan...

In this paper, we present a novel approach for segmenting planar regions in Digital Surface Models (DSMs) by adapting the Segment Anything Model (SAM), an open-source framework. Our approach specifically tailors SAM to recognize planar regions within given building footprints, employing the Low-Rank Adaptation (LoRA) technique. This adaptation bene...

Throughout the evolution of machine learning, the size of models has steadily increased as researchers strive for higher accuracy by adding more layers. This escalation in model complexity necessitates enhanced hardware capabilities. Today, state-of-the-art machine learning models have become so large that effectively training them requires substan...

Anomaly Detection is an important topic in various application areas, including image analysis and network intrusion detection. The Reed–Xiaoli (RX) detector is an efficient and accurate anomaly detector that can be used if analyzed data is Gaussian distributed. However, in the real-world, data is rarely Gaussian distributed. For nonlinear data, ke...

In this study, the performance of semi-global methods for satellite stereo image matching was investigated. The performance of semi-global methods for the well-known Middlebury and KITTI sets has been the subject of many studies. Applications using satellite images may have different difficulties compared to close-range images. RVLSatStereo was use...

In the last decade, Deep Learning is applied in a wide range of problems with tremendous success. Large data, increased computational resources, and theoretical improvements are main reasons for this success. As the dataset grows, the realworld is better represented, allows developing a model that can generalize. However, creating a labeled dataset...

Modeling with ground-truth is a field of study that will take up much more space in the future with the developing technology in areas such as cartography, transportation, communication, exploration, search and rescue, smart city. Ground-truth data are important needs for researchers working in the field of remote sensing. Data collection and usage...

In the last decade, Deep Learning is applied in a wide range of problems with tremendous success. This success mainly comes from large data availability, increased computational power, and theoretical improvements in the training phase. As the dataset grows, the real world is better represented, making it possible to develop a model that can genera...

In this work we propose a method to find anomalous changes in remote sensing images based on the chronochrome approach. A regressor between images is used to discover the most {\em influential points} in the observed data. Typically, the pixels with largest residuals are decided to be anomalous changes. In order to find the anomalous pixels we cons...

This work tackles the target detection problem through the well-known global RX method. The RX method models the clutter as a multivariate Gaussian distribution, and has been extended to nonlinear distributions using kernel methods. While the kernel RX can cope with complex clutters, it requires a considerable amount of computational resources as t...

We here introduce an automatic Digital Terrain Model (DTM) extraction method. The proposed sparsity-driven DTM extractor (SD-DTM) takes a high-resolution Digital Surface Model (DSM) as an input and constructs a high-resolution DTM using the variational framework. To obtain an accurate DTM, an iterative approach is proposed for the minimization of t...

To be presented.

To be presented.

Detecting anomalous changes in remote sensing images is a challenging problem where many approaches and techniques have been presented so far. We rely on the standard field of multivariate statistics of diagnostic measures which is concerned about the characterization of distributions, detection of anomalies, extremes events and changes. One useful...

Visual and automatic analyses using synthetic aperture radar (SAR) images are challenging because of inherently formed speckle noise. Thus, reducing speckle noise in SAR images is an important research area for SAR image analysis. During speckle noise reduction, homogeneous regions should be smoothed while details such as edges and point scatterers...

In this study, we propose a new segmentation approach for vessel structures based on fluid flow. Vessels are medium for transportation of blood which makes segmentation using fluid flow physically plausible. So, streaming fluid regions can capture smooth and thin structures. It also provides robustness to radiofrequency field inhomogeneity and nois...

Yabancı otların otomatik tanınması hassas tarım (precision agriculture) için önemli adımlardan biridir. Ayrıca, otomasyon sistemlerinin önemli bir bileşeni olan tarım robotlarında gerçek zamanlı olarak hedef tanıma yapılması kritik önemdedir. Bu çalışmada yabancı otların otomatik tanınması için 4 derin öğrenme mimarisinin performansı karşılaştırılm...

In this study, a novel approach which combines the advantages of Total Variation based Sparsity Driven Despeckling (SDD-QL) and Nonlocal Means is proposed in order to improve the despeckling quality. Both SDD-QL and Nonlocal Means have advantages and disadvantages on image despeckling. SDD-QL performs really well on homogeneous areas and quite fast...

Noise is an unwanted signal resides in images that deteriorates the crucial information and structures in images. In this study, the advantages of Non-local Means filter and Total Variation based Sparsity Driven Despeckling with Quadratic Linear term is combined in a single cost function. NL means is used on texture areas and SDD-QL is used on homo...

This work tackles the target detection problem through the well-known global RX method. The RX method models the clutter as a multivariate Gaussian distribution, and has been extended to nonlinear distributions using kernel methods. While the kernel RX can cope with complex clutters, it requires a considerable amount of computational resources as t...

In this work we propose a method to nd anomalous changes in remote sensing images based on the chronochrome approach. A regressor between images is used to discover the most inuential points in the observed data. Typically, the pixels with largest residuals are decided to be anomalous changes. In order to nd the anomalous pixels we consider the Coo...

We here introduce an automatic Digital Terrain Model (DTM) extraction method. The proposed sparsity-driven DTM extractor (SD-DTM) takes a high-resolution Digital Surface Model (DSM) as an input and constructs a high-resolution DTM using the variational framework. To obtain an accurate DTM, an iterative approach is proposed for the minimization of t...

In this study, we propose a correlation based variational change detection (CVCD) method for elevation models. In essence, CVCD aims to produce smooth change maps while preserving the details of the terrain by minimizing a varia-tional cost function. Proposed cost function is constructed with a novel data fidelity term using normalized correlation...

Speckle noise, inherent in synthetic aperture radar (SAR) images, degrades the performance of the various SAR image analysis tasks. Thus, speckle noise reduction is a critical preprocessing step for smoothing homogeneous regions while preserving details. Therefore, recently, SDD-QL method was proposed which is a variational approach where L1-norm t...

In this study, a novel sparsity-driven weighted ensemble classifier (SDWEC) that improves classification accuracy and minimizes the number of classifiers is proposed. Using pre-trained classifiers, an ensemble in which base classifiers votes according to assigned weights is formed. These assigned weights directly affect classifier accuracy. In the...

Speckle noise, inherent in synthetic aperture radar (SAR) images, degrades the performance of the various SAR image analysis tasks. Thus, speckle noise reduction is a critical preprocessing step for smoothing homogeneous regions while preserving details. This letter proposes a variational despeckling approach where L1-norm total variation regulariz...

As the first step of automatic image interpretation systems, automatic detection of targets should be accurate and fast. For Synthetic Aperture Radar (SAR) images, Constant False Alarm Rate (CFAR) is the most popular framework used for target detection. In CFAR, modeling of the clutter is crucial since the decision threshold is calculated based on...

As the first step of automatic image interpretation systems, automatic detection of the targets should be accurate and fast. Constant False Alarm Rate (CFAR) is the most popular target detection framework for Synthetic Aperture Radar (SAR) images. For CFAR, modeling of the clutter is crucial since the decision threshold is calculated based on this...

Speckle noise formed in Synthetic Aperture Radar (SAR) images makes visual and automatic analyses complicated. Thus, reducing speckle noise in homogeneous regions while preserving features such as edges and point scatterers is important as a pre-processing step. Although SAR images predominantly contains multiplicative noise, it also contains low a...

In this study, Sparsity-driven Change Detection (SDCD) method, which has been proposed for detecting changes in multitemporal synthetic aperture radar (SAR) images, is parallelized to reduce the execution time. Parallelization of the SDCD is realized using OpenMP on CPU and CUDA on GPU. Execution speed of the parallelized SDCD is shown on real-worl...

Cranioplasty is a neurosurgical operation for repairing cranial defects that have occurred in a previous operation or trauma. Various methods have been presented for cranioplasty from past to present. In computer-aided design based methods, quality of an implant depends on operator’s talent. In mathematical model based methods, such as curve-fittin...

Sentetik Açıklıklı Radar (SAR) görüntülerinde otomatik hedef tespiti önemli bir çalışma alanı olarak göze çarpmaktadır. Bu bağlamda ele alınan hedeflerden bir tanesi olan gemi tespiti üzerinde gerçekleştirilen çalışmaların sayısı son yıllarda önemli ölçüde artmıştır. Literatürde önerilen çalışmalar genellikle açık denizdeki bağımsız gemiler üzerind...

Değişiklik tespiti, dünya yüzeyini gözlemlemek amacıyla kullanılan oldukça önemli fakat zorlu bir araştırma konusudur. Araştırmacılar değişiklik tespiti yöntemini gerçekleştirirken görüntülerin örtüştürülmemesi, görüntülerin gürültülü olması ve gözle fark edilmeyen farklılıklardan kaynaklanan bazı sorunları çözmek için algoritmalar uygulamışlardır....

Hiperspektral görüntüler farklı dalga boylarında ait oldukları materyaller ile ilgili spektral bilgiler taşıyan ardışık yüzlerce banttan oluşmaktadırlar. Bununla birlikte, atmosferik etkiler ve sensör yapısı gibi pratik faktörlerden dolayı bazı spektral bantlar yüksek seviyede gürültü içermektedir. Bu bildiride, gürültülü bantları otomatik ve etkin...

In this letter, a novel weighted ensemble classifier is proposed that improves classification accuracy and minimizes the number of classifiers. Ensemble weight finding problem is modeled as a cost function with following terms: (a) a data fidelity term aiming to decrease misclassification rate, (b) a sparsity term aiming to decrease the number of c...

In this letter, a method for detecting changes in multitemporal synthetic aperture radar (SAR) images by minimizing a novel cost function is proposed. This cost function constructed with a log-ratio based data fidelity terms and ℓ1-norm based total variation (TV) regularization term. Log-ratio terms model the changes between the two SAR images wher...

Digital surface models (DSM) are crucial for applications such as surface deformation analysis using synthetic aperture radar (SAR) interferometry, automatic target recognition, ortorectification of airborne and satellite images, and generation of digital terrain model (DTM). SAR interferometry, SAR radargrammetry, electro optic (EO) photogrammetry...

Speckle noise which is inherent to Synthetic Aperture Radar (SAR) imaging makes it difficult to detect targets and recognize spatial patterns on earth. Thus, despeckling is critical and used as a preprocessing step for smoothing homogeneous regions while preserving features such as edges and point scatterers. In this study, a low-memory version of...

Segmentation is widely used for determining tumor and other lesions and classifying tissues for various analysis purposes in medical images. However, being an ill-posed problem, there is no single segmentation method which can perform successfully for all kind of data. In this study, a novel total variation (TV) based skull segmentation method is p...

SAR remote sensing systems provide high-resolution images of the earth's surface as seen from airborne or satellite platforms. SAR imaging has a growing interest in remote sensing applications in many fields with the ability to generate high-resolution radar image regardless of weather conditions and sunlight illumination. However, due to coherent...

Modeling the earth surface and ground is a crucial task for various remote sensing applications such as map generation, 3d model generation, vegetation extraction, building extraction, bridge extraction, and river analysis. However, obtaining accurate earth surface and ground is a challenging task since sensor retrieves measurements from both groun...

Sabanci University 2 , Istanbul, Turkey As an active sensing system, Synthetic Aperture Radar (SAR) has the advantage of working in all weather conditions during day and night and it has also capability to penetrate the cloud coverage. Due to these properties, SAR systems can be considered as a vital source for SAR image analysis applications such...

Automatic target recognition (ATR) using synthetic aperture radar (SAR) images is a major research area of remote sensing community. In this context, ship recognition comes as one of the important and challenging task. Ship recognition is also very important for various civil and military applications and the number of studies in this direction has...

Mustafa ERGÜL, Çağdaş BAK, Emre AKYILMAZ, Fatih NAR, Nigar ŞEN SDT Space and Defense Technologies, Ankara, Turkey With the recent developments on optical sensing technology, hyperspectral imagery (HSI) has reached the ability to record rich spectral and spatial information of the observed scene. The tremendous amount of spatial and spectral informa...

Speckle noise inherent in synthetic aperture radar (SAR) images seriously affects the result of various SAR image processing tasks such as edge detection and segmentation. Thus, speckle reduction is critical and is used as a preprocessing step for smoothing homogeneous regions while preserving features such as edges and point scatterers. Although s...

Land-cover classification in Synthetic Aperture Radar (SAR) images has significance in both civil and military remote sensing applications. Accurate classification is a challenging problem due to variety of natural and man-made objects, seasonal changes at acquisition time, and diversity of image reconstruction algorithms.. In this study, Feature P...

Speckle noise which is inherent to Synthetic Aperture Radar (SAR) imaging obstructs various image exploitation tasks such as edge detection, segmentation, change detection, and target recognition. Speckle reduction is generally used as a first step which has to smooth out homogeneous regions while preserving edges and point scatterers. In remote se...

Hyperspectral imagery consists of hundreds of successive bands that carry spectral information about the underlying materials at various wavelengths. However, due to practical factors such as atmospheric effects and sensor characteristics, some spectral bands contain high amounts of noise. In this paper, an effective information-theoretic algorithm...

Speckle noise which is inherent to Synthetic Aperture Radar (SAR) imaging obstructs various image exploitation tasks such as edge detection, segmentation, change detection, and target recognition. Therefore, speckle reduction is generally used as a first step which has to smooth out homogeneous regions while preserving edges and point scatterers. T...

Synthetic Aperture Radar (SAR) images contain high amount of speckle noise which causes edge detection, shape analysis, classification, segmentation, change detection and target recognition tasks become more difficult. To overcome such difficulties, smoothing of homogenous regions while preserving point scatterers and edges during speckle reduction...

Cranioplasty is a surgical operation to repair hole or defects on skull. 3 dimensional computed tomography (CT) images are used for automatic determination of the shape of implant which is used for repairing defect. The designing implant by mathematical model and manufacturing it before operation lowers the operation cost. In this paper, previous s...

Vessel segmentation is important for many clinical applications, such as the diagnosis of vascular diseases, the planning of surgery, or the monitoring of the progress of disease. Although various approaches have been proposed to segment vessel structures from 3-dimensional medical images, to the best of our knowledge, there has been no known techn...

In this paper, we propose a change detection approach based on nonlinear scale space analysis of change images for robust detection of various changes incurred by natural phenomena and/or human activities in Synthetic Aperture Radar (SAR) images using Maximally Stable Extremal Regions (MSERs). To achieve this, a variant of the log-ratio image of mu...

Ship detection from synthetic aperture radar (SAR) images is important for various automatic target recognition (ATR) tasks. Although the ships in offshore areas can be easily detected, the ones near the shores or close to each other are difficult to detect. Furthermore, segmentation and classification of such ships is extremely difficult. In this...

Extraction of curvilinear features from synthetic aperture radar (SAR) images is important for automatic recognition of various targets, such as fences, surrounding the buildings. The bright pixels which constitute curvilinear features in SAR images are usually disrupted and also degraded by high amount of speckle noise which makes extraction of su...

Enhancement of retinal images supports both manual examination of anomalies by medical experts and also automatic segmentation, registration, and classification tasks. Three kind of enhancements; brightness adjustment, contrast enhancement and noise reduction needs to target different tissues in a different way. For example, contrast enhancement ha...

Sentetik açıklıklı radar (SAR) görüntülerinde otomatik hedef tespiti yöntemleri görüntünün çözünürlüğüne, hedefin büyüklüğüne, parazit yankı karmaşıklığına ve benek gürültü seviyesine duyarlıdır. Gürbüz bir hedef tespiti yönteminin ise bu tür etkenlere daha az duyarlı olması istenir. Önerilen yöntem görüntünün öznitelik korumalı benek gürültü arınd...

Automatic target detection (ATD) methods for synthetic aperture radar (SAR) imagery are sensitive to image resolution, target size, clutter complexity, and speckle noise level. However, a robust ATD method needs to be less sensitive to the above factors. In this study, a constant false alarm rate (CFAR) based method is proposed which can perform ta...

We consider the problem of despeckling synthetic aperture radar (SAR) images and propose an approach we call feature preserving despeckling (FPD). FPD is obtained through the adoption of a regularized SAR image reconstruction algorithm for the despeckling problem. FPD performs smoothing of homogeneous regions while preserving strong scatterers as w...

Automatic target detection methods for synthetic aperture radar (SAR) images are sensitive to image resolution, size of the target to be detected, clutter complexity, and speckle noise level. A robust automatic target detection method needs to be less sensitive to the above factors. In this study, a constant false alarm rate (CFAR) based automatic...

In this study a generic framework for the parallel implementation of Simulated Annealing (SA) method is proposed. In proposed method search space is spanned by local SA search agents (SASAs) where optimum distribution of agents in search space is obtained using k-means clustering algorithm. Clusters are formed dynamically based on spatial position...

Optimum scheduling of turn of duty for pharmacies is a crucial problem for Ankara Pharmacist Association (APA). Six requirements for optimum scheduling of turn of duty are (a) choosing sentinel pharmacies for each day so that coverage area is maximized and making distance between each sentinel pharmacy as maximum, (b) producing fair scheduling so t...

Cortical renal (kidney) scintigraphy images are 2D images (256x256) acquired in three projection angles (posterior, right-posterior-oblique and left-posterior-oblique). These images are used by nuclear medicine specialists to examine the functional morphology of kidney parenchyma. The main visual features examined in reading the images are: size, l...

Increasing amount of image data raises the importance of content based query systems. Increasing hardware capacity and improving methods makes development of such systems more feasible. In this study, we aim to develop a content based image retrieval for renal (kidney) scintigraphy images. For this purpose, problem analysis, literature survey, data...

Subtraction of ictal and interictal single photon emission computed tomography (SPECT) images is known to be successful in localizing the seizure focus in the pre-surgical evaluation of patients with partial epilepsy. A computer-aided methods for producing subtraction ictal SPECT co-registered to the magnetic resonance image (MRI) (the SISCOM metho...

Subtraction of ictal and interictal SPECT images is known to be successful in localizing the seizure focus in presurgical evaluation of patients with partial epilepsy. Computer-aided method for producing subtraction ictal SPECT coregistered to MRI (SISCOM method) is commonly used. There are two registrations involved in SISCOM: between the ictalint...

Subtraction of ictal and interictal single-photon emission computed tomography (SPECT) images is known to be successful in localizing the seizure focus in presurgical evaluation of patients with partial epilepsy. The subtracted images are also aligned with a patient’s highresolution magnetic resonance image (MRI) and fused to identify the anatomica...

An example based rendering (EBR) method based on generalization and localization that uses artificial neural networks (ANN) and k-Nearest Neighbor (k-NN) is proposed. The method involves learning phase and application phase, which means that once a transformation filter is learned, it can be applied to any other image. In learning phase, error back...