Akademik Görevleri

  • Present 2016

    Dr. Öğr. Üyesi

    Karabük Üniversitesi, Bilgisayar Mühendisliği

  • 2018 2017

    Misafir Öğr. Üyesi

    Purdue Üniversitesi, Elektrik ve Bilgisayar Mühendisliği, Indiana, ABD

  • 2016 2008

    Arş. Gör.

    Karabük Üniversitesi, Bilgisayar Mühendisliği

Eğitim Bilgileri

  • Ph.D. 2015

    Bilgisayar Mühendisliği ABD

    Karabük Üniversitesi, Fen Bilimleri Enstitüsü

  • M.Sc. 2011

    Bilgisayar Mühendisliği ABD

    Karabük Üniversitesi, Fen Bilimleri Enstitüsü

  • B.Sc.2008

    Bilgisayar Mühendisliği Bölümü

    Yıldız Teknik Üniversitesi, Elektrik Elektronik Fakültesi

Projeler

  • image

    KBÜ BAP Projesi

    "Büyük Verinin Makine Öğrenmesi Yöntemleri ile Apache Spark Teknolojisi Kullanılarak Sınıflandırılması", Proje No: KBÜBAP-17-YL-072, Proje Yürütücüsü, Ocak 2017 - Kasım 2017

    Bu çalışmada, teknolojinin ve internetin hızla gelişmekte olduğu bilgi çağında verilerin üretimi, depolanması, analiz edilmesi ve analiz sonuçlarının büyük bir değere sahip olduğundan dolayı büyük veri üzerinde çalışılmıştır. Büyük veri üzerinde sınıflandırma ve kümeleme işlemleri zaman alıcı olabilmektedir. Bu çalışmada, büyük verinin işlenmesi ve analiz edilmesi için geliştirilen Apache Spark teknolojisi kullanılarak farklı büyük veriler üzerinde sınıflandırma, kümeleme ve aykırı değer algılama işlemlerinin yapılması amaçlanmıştır.

    Bu amaçla, makine öğrenmesi algoritmalarını içeren Apache Spark’ın MLlib kütüphanesinden faydalanılmıştır. Apache Spark teknolojisini kullanarak hataya dayanıklı, güvenilir, tutarlı ve hızlı sınıflandırma ve kümeleme işlemi gerçekleştirmesi amaçlanmaktadır. Bu çalışmada kullanılan MLlib kütüphanesinde yer alan Naïve Bayes, K-means ve Gaussian Mixture yöntemleri ile büyük verilerin başarılı bir şekilde analiz edilmesi sağlanmış algoritmaların çalışma süreleri farklı veri boyutları kullanılarak tespit edilmiştir. K-means kümeleme algoritmasının uygulaması Spark Standalone modda, 1 master ile 1 master 3 worker şeklinde çalıştırılıp çalışma süreleri tespit edilmiştir.

  • image

    YBÜ BAP Projesi

    "Görüntüler Üzerindeki Gürültüyü Gidermede Hibrit Yöntem Geliştirilmesi", Proje No: YBU-BAP-590, Araştırmacı, Ağustos 2013 - Ağustos 2015

    Bu çalışmada, algoritma çalışma süresi kısa tutularak gürültü azaltma performansının geliştirilmesi için l0-norm, l1-norm ve l2-normun avantajları ve yerel olmayan ortalama filtrenin doku koruma özelliği bir araya getirilerek yeni bir yöntem geliştirilmektedir. Önerilen yöntem görüntüdeki kenar ve noktasal saçıcıların bozulmasını önleyerek homojen bölgelerin yumuşatılmasını sağlamaktadır. Bu alanda yapılan çalışmalar incelendiğinde l0-norm, l1-norm ve l2-normun değişik uygulama senaryolarında birbirlerine göre avantajları olduğu görülmektedir. Görüntülerdeki her bir piksel için bu normlardan uygun olanı kullanılırsa daha iyi bir gürültü azaltma sağlanacaktır. Ayrıca yerel olmayan ortalama filtrenin kullanımı ile de görüntü üzerindeki dokular daha iyi korunabilecektir.

    Bu yöntemleri bir arada kullanabilmek için yöntemlerin sahip olacağı farklı ağırlık değerlerinden oluşan bir maliyet fonksiyonu tanımlanmaktadır. Önerilen maliyet fonksiyonunun eşlenik gradyan yöntemleri kullanılarak minimize edilmesiyle gürültü azaltma gerçekleştirilmektedir. Önerilen yöntem etkin bellek kullanımı sağlamaktadır ve yöntemin tüm adımlarının CPU’da OpenMP ve GPU’da CUDA kullanılarak paralelleştirilmesi sağlanmaktadır. Benek gürültü azaltma performansı, çalışma süresi ve bellek kullanımı sentetik ve gerçek SAR görüntüleri üzerinde test edilmiştir.

  • image

    Desteklenen Lisans Bitirme Projeleri

    ∴ KBÜ Lisans BAP Projeleri

    ∴ Tübitak 2241-Sanayi Odaklı Lisans Bitirme Projeleri

    ∴ Tübitak 2209-Üniversite Araştırma Projeleri

    WeWantEducation: Öğrenciler Arası Eğitim Ağının Mobil Uygulama İle Sağlanması, Karabük Üniversitesi Lisans Öğrencisi Araştırma Projesi, Proje Yürütücüsü, Proje No: KBÜBAP-17-LÖAP-127, 2017 (Proje Öğrencisi: Cahit Berkay Kazangirler)

    SECUREYELLOW: Ticari Taksilerde Yüz Tanıma Sistemi İle Suçlu Analizi, Karabük Üniversitesi Lisans Öğrencisi Araştırma Projesi, Proje Yürütücüsü, Proje No: KBÜBAP-17-LÖAP-122, 2017 (Proje Öğrencileri: Sinem Erdoğan, Halil İbrahim Çopurkuyu)

    Makine Öğrenmesi ve Veri Madenciliği Yöntemleri ile Büyük Veri Analizi, Karabük Üniversitesi Lisans Öğrencisi Araştırma Projesi, Proje Yürütücüsü, Proje No: KBÜBAP-17-LÖAP-123, 2017 (Proje Öğrencisi: İskender Ülgen Oğul)

     

    Maden Araç Takip Sistemi, Tübitak 2241, Proje Danışmanı, 2017 (Proje Öğrencileri: Semih Celal, Muhammet Taha Aydın)

     

    Görüntü İşleme ve Veri Madenciliği Yöntemleri Kullanılarak Görüntü Analizi (Gözcü++), Tübitak 2209, Proje Danışmanı, 2017 (Proje Öğrencileri: Mustafa Bulut, Sercan Çolak)

    Renklerdeki Müzik, Tübitak 2209, Proje Danışmanı, 2017 (Proje Öğrencileri: Kübra Özdemir, Mustafa Mert Gün)

    Kanveren: Acil Kan İhtiyaci İçin Mobil Çözüm, Tübitak 2209, Proje Danışmanı, 2017 (Proje Öğrencileri: Onur Aydin, Hasan Akgül)

  • image

    Tübitak Projesi

    "Bilgisayar Kontrollü Ağaç Malzeme Yanma Düzeneği", Proje No: 110O548, Bursiyer, 2011 - 2013

    Bu çalışmada amaç, bilgisayar kontrollü ağaç malzeme yanma düzeneği hazırlanarak yanma performansının bulanık mantık ile izlenmesidir. Bu sayede manuel ölçme ve insan kaynaklı hataların en aza indirilmesi, yanma sonucu elde edilen parametrelerin ve yanma sürecinin izlenmesi gerçekleştirilmektedir. Yanma esnasında yapılan ölçümlerin bilgisayar ortamına aktarılarak gerçek zamanlı olarak kaydedilmesi sağlanmaktadır. Yanma deneyi süresince ölçülen değerler üzerine bulanık çıkarım uygulanarak deneyin gerçek zamanlı performansı değerlendirilmekte ve yanma işlemi kontrol altında tutulmaktadır.

    Yanma işlemi boyunca ve bitiminde elde edilen verilerin işlenmesi ve analizlerin yapılması yine bilgisayar ortamında gerçekleştirilmekte, elde edilen veriler tablo ve grafiklerle sunulmaktadır. Ölçüm sonucu kaydedilen tüm verilerin daha sonra yapılacak yanma çalışmalarında faydalanılmak üzere depolanması sağlanmaktadır. Bu sayede ileride yapılacak çalışmaların doğruluğu deneysel tecrübelerle kontrol edilebilecektir. Bu zamana kadar yapılan yanma çalışmaları bilgisayar kullanılmaksızın verilerin gözlemlerle okunması ve kaydedilmesi şeklindedir. Bu durum insan ve ölçme kaynaklı hataları beraberinde getirmektedir.

    Bu çalışmada ise hataların tespit edilmesi ve deney performansının izlenmesi gerçek zamanlı olarak sağlanmaktadır. Hataların önceden tahmin edilmesi özellikle zaman ve iş gücünden tasarruf sağlamaktadır. Bu sayede düzenek, ağaç malzeme koruma teknolojisi ve ağaç malzemenin yanma özelliklerinin belirlenmesi alanında kullanılabilme özelliklerine sahip olmaktadır. Tarihi ahşap yapıların yanmaya karşı korunmasıyla ilgili yapılacak çalışmalara imkan tanınmaktadır. Tarihi ahşap evlerin yanma tehlikesine karşı korunabilmesi de tarihi kültürümüzün sürdürülebilirliğine katkı sağlayacaktır.

Filtreleme:

Yıla Göre Sırala:

Enhanced multidimensional field embedding method by potential fields for hyperspectral image classification and visualisation

C. Ozcan and O. Ersoy
SCI Makale Electronics Letters, DOI:10.1049/el.2018.0676 , 10 April 2018

Abstract

Multidimensional field embedding methods have been demonstrated to effectively characterise spectral signatures in hyperspectral images. However, high-dimensional data composed of a number of classes presents challenges to the existing embedding methods. This Letter proposes an enhanced multidimensional field embedding algorithm based on the force field formulation. The comparative performance of the proposed algorithm is evaluated in the classification and visualisation of commonly used hyperspectral images. Experimental results demonstrate its superiority over previously used field embedding techniques.

Sparsity-Driven Despeckling for SAR Images (Matlab Codes)

C. Ozcan, B. Sen and F. Nar
SCI Makale IEEE Trans. Geosci. Remote Sens., 13(1):115-119, Nov. 2015

Abstract

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 state-of-the-art methods provide better despeckling compared with conventional methods, their resource consumption is higher. In this letter, a sparsity-driven total-variation (TV) approach employing l0-norm, fractional norm, or l1-norm to smooth homogeneous regions with minimal degradation in edges and point scatterers is proposed. Proposed method, sparsity-driven despeckling (SDD), is capable of using different norms controlled by a single parameter and provides better or similar despeckling compared with the state-of-the-art methods with shorter execution times. Despeckling performance and execution time of the SDD are shown using synthetic and real-world SAR images.

Cropped Quad-Tree Based Solid Object Colouring With Cuda

Çavuşoğlu A., Şen B., Özcan C. and Görgünoğlu S.
Ulusal Makale Mathematical and Computational Applications, 18(3):301-312, 2013

Özet

In this study, surfaces of solid objects are coloured with Cropped Quad-Tree method utilizing GPU computing optimization. There are numerous methods used in solid object colouring. When the studies carried out in different fields are taken into consideration, it is seen that quad-tree method displays a prominent position in terms of speed and performance. Cropped quad-tree is obtained as a result of the developments seen with the frequent use of this method in the field of computer sciences. Two different versions of algorithm which operate recursively on CPU and at the same time which use GPU computing optimization are used in this study. Besides, OpenGL is used for graphics drawing process. Within the setting of the study, results are obtained via CPU and GPU’s, at first using Quad-Tree method and then Cropped Quad-Tree method. It is observed that GPU computing is obviously faster than CPU computing and Cropped Quad-Tree method produces rapid results compared to Quad-Tree method as a result of performance. GPU computing method boosted approximately performance by up to 20 times compared to only CPU usage; furthermore, cropped quad-tree method boosted approximately performance of algorithm by up to 25 times on average dependent on screen and object size.

Classification of SAR Image Patches with Apache Spark Using GLCM Texture Features

Caner Ozcan, Okan Ersoy, Iskender Ulgen Ogul
Uluslararası Bildiri International Conference on Advanced Technologies, 3rd World Conference on Big Data (WC-BIGDATA 2018), 28-30 Apr., Izmir, Turkey, 2018

Özet

Synthetic-aperture radar (SAR) remote sensing systems allow images of the visible parts of the earth's surface to be acquired with high resolution. Due to its ability to produce high-resolution radar images regardless of weather conditions and sunlight illumination, it has a growing interest in the remote sensing community. SAR image classification plays an important role in the analysis and interpretation of SAR images, with a large number of applications in many areas. However, classification of high-resolution SAR images is a time-consuming process due to their sizes. In this study, high speed SAR image patch classification by using Apache Spark framework is discussed. Spark is an open source library which offers unique advantages in-memory capabilities, and is well suited for big datasets. We discuss successful image classification results obtained by using the Naive Bayes method with the Spark MLlib library. The Naive Bayes method is a supervised learning algorithm based on applying Bayes’ theorem to the input feature vectors, which are obtained from each SAR image patch by using gray-level co-occurrence matrix (GLCM). Public TerraSAR-X high resolution images were used as real-world SAR images in the experimental studies.

Automatic Segmentation and Labelling of 3D Human Activities

Rafet Durgut, Caner Ozcan, Oguz Findik
Uluslararası Bildiri International Conference on Advanced Technologies, Computer Engineering and Science (ICATCES), 11-13 May, Karabuk, Turkey, 2018

Özet

Recognition and interpretation of human activities are very interesting and hot topics that are frequently studied in the field of computer vision. Especially with the advent and development of the Microsoft Kinect depth sensors, the expansion of the study field has gained momentum in the positive direction. Thanks to RGBD cameras, which also provide depth information in addition to the RGB image, researchers benefit from many advantages in terms of privacy, accuracy and precision. In this study, automatic segmentation of repeated 3D human activity is proposed. A public dataset containing the repeated action sequences are recorded using the RGBD camera. The action sequence in this dataset includes similar and different action information. In order to identify and label each action in sequence, it is necessary to perform the segmentation process. To be able to perform a successful segmentation process, the data must be pre-processed to remove noise. For this purpose, a total variation based noise removal method is used. Human action recognition and detailed error analysis can be performed through the segments derived from the output of this work.

Implementation of NURBS curves on the LCD touch screen using FPGA

Yasin Ozturk, Caner Ozcan
Uluslararası Bildiri International Conference on Advanced Technologies, Computer Engineering and Science (ICATCES), 11-13 May, Karabuk, Turkey, 2018

Özet

- Computer-aided curve and surface design models are used to design surfaces that do not have a particular shape. Since these surfaces cannot be expressed by known mathematical functions, curve and surface modeling methods such as Bezier, B-Spline and Non-Uniform Rational B-Spline (NURBS) have been developed. In this study, an application is developed that allows the NURBS curve algorithm, which is used as a method in 3D modeling, to be computed using Field Programmable Gate Array (FPGA) with parallel processing capability and displayed on the LCD touch screen. The B-Spline basis function which NURBS based on are sampled mathematically. The parametric values used in the display of the NURBS curves and surfaces are presented on the LCD with user interaction. Parametric values such as control points, knot vector, and weight vector are separately sampled for curves and mathematical solutions are given. The NURBS curve for user-defined control points can be displayed in real time on the screen. In addition, an implementation of the NURBS algorithm in the Visual Studio platform was developed and obtained results were compared.

Stream Text Data Analysis on Twitter Using Apache Spark Streaming

Ozlem Hakdagli, Caner Ozcan, Iskender Ulgen Ogul
Uluslararası Bildiri IEEE Signal Processing and Communications Applications Conference (SIU), 26th, Page(s): -, Izmir, Turkey, 2-5 May, 2018

Özet

With today's developing technology, people's access to information and its production have reached a very fast level. These generated and obtained information are instantly created, entered into data systems and updated. Sources of streaming data can be transformed into valuable analysis results when they are handled with targeted methods. In this study, a text data field is determined to perform analysis on instantaneous generated data and Twitter, the richest platform for instant text data, is used. Twitter instantly generates a variety of data in large quantities and it presents it as open source using an API. A machine learning framework Apache Spark's stream analysis environment is used to analyze these resources. Situation analysis was performed using Support Vector Machine, Decision Trees and Logistic Regression algorithms presented under this environment. The results are presented in tables.

Real-Time Automation System for Classifying Fruits with Image Processing

Ismail Burak Akinci, Caner Ozcan
Uluslararası Bildiri International Symposium on Industry 4.0 and Applications, 12-14 Oct., Karabuk, Turkey, 2017

Özet

Image processing system, which is inspired by the vision of individuals, is now used in many sciences. Image processing which involves transferring and processing images to a computer using different methods and tools has resulted in the acquisition of a new image or interpretation of the current image. Image processing systems in the industry have also contributed to increased production requirements. At this point, the areas where image processing systems are used have been utilized with minimum cost and maximum efficiency principle. In this study, an automation system has been realized which classifies the fruits in the video images according to their dimensions in real-time with the EmguCV graphic library which is used as a method in image processing models. The filter models that form the basis of the EmguCV library are illustrated in this study. Mathematical solutions of the image processing are given separately for image smoothing and edge detection. Obtained results can be displayed in the application using OpenGL from net Framework classes. The application is built on Visual Studio 2012, OpenGL 3.0 and Windows 7 platforms.

Extraction of Greenhouse Areas with Image Processing Methods in Karabuk Province

Yildirim, M. Z. and Ozcan, C.
Uluslararası Bildiri Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W6, 107-109, 2017

Özet

Greenhouses provide the environmental conditions to be controlled and regulated as desired while allowing agricultural products to be produced without being affected by external environmental conditions. High quality and a wide variety of agricultural products can be produced throughout the year. In addition, mapping and detection of these areas has great importance in terms of factors such as yield analysis, natural resource management and environmental impact. Various remote sensing techniques are currently available for extraction of greenhouse areas. These techniques are based on the automatic detection and interpretation of objects on remotely sensed images. In this study, greenhouse areas were determined from optical images obtained from Landsat. The study was carried out in the greenhouse areas in Karabuk province. The obtained results are presented with figures and tables.

Fast Text Classification with Naive Bayes Method on Apache Spark

Iskender Ulgen Ogul, Caner Ozcan, Ozlem Hakdagli
Uluslararası Bildiri IEEE Signal Processing and Communications Applications Conference (SIU), 2017 25th, Page(s): -, Antalya, Turkey, 15-18 May, 2017

Özet

The increase in the number of devices and users online with the transition of Internet of Things (IoT), increases the amount of large data exponentially. Classification of ascending data, deletion of irrelevant data, and meaning extraction have reached vital importance in today's standards. Analysis can be done in various variations such as Classification of text on text data, analysis of spam, personality analysis. In this study, fast text classification was performed with machine learning on Apache Spark using the Naive Bayes method. Spark architecture uses a distributed in-memory data collection instead of a distributed data structure presented in Hadoop architecture to provide fast storage and analysis of data. Analyzes were made on the interpretation data of the Reddit which is open source social news site by using the Naive Bayes method. The results are presented in tables and graphs.

Sparsity-Driven Despeckling Method with Low Memory Usage

Ozcan, Caner; Sen, Baha; Nar, Fatih
Uluslararası BildiriIEEE Signal Processing and Communications Applications Conference (SIU), 2016 24nd, Page(s): 1329-1332, Zonguldak, Turkey, 16-19 May, 2016

Özet

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 the previously proposed sparsity-driven despeckling (SDD) method is proposed. All steps of the method are parallelized using OpenMP on CPU and CUDA on GPU. Execution time and despeckling performance are shown using real-world SAR images.

Total variation based 3D skull segmentation

Atasoy, Ferhat; Sen, Baha; Nar, Fatih; Ozcan, Caner; Bozkurt, Ismail
Uluslararası BildiriIEEE Signal Processing and Communications Applications Conference (SIU), 2016 24nd, Page(s): 1353-1356, Zonguldak, Turkey, 16-19 May, 2016

Özet

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 proposed. Skull segmentation performance of the proposed method is shown using computed tomography (CT) images.

GPU efficient SAR image despeckling using mixed norms

Caner Ozcan, Baha Sen, Fatih Nar
Uluslararası BildiriProc. SPIE 9247: High-Performance Computing in Remote Sensing IV, Amsterdam, Netherlands, 2014

Özet

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. Traditional speckle reduction methods are fast and their memory consumption is insignificant. However, they are either good at smoothing homogeneous regions or preserving edges and point scatterers. State of the art despeckling methods are proposed to overcome this trade-off. However, they introduce another trade-off between denoising quality and resource consumption, thereby higher denoising quality requires higher computational load and/or memory consumption. In this paper, a local pixel-based total variation (TV) approach is proposed, which combines l2-norm and l1-norm in order to improve despeckling quality while keeping execution times reasonably short. Pixel-based approach allows efficient computation model with relatively low memory consumption. Their parallel implementations are also more efficient comparing to global TV approaches which generally require numerical solution of sparse linear systems. However, pixel-based approaches are trapped to local minima frequently hence despeckling quality is worse comparing to global TV approaches. Proposed method, namely mixed norm despeckling (MND), combines l2-norm and l1-norm in order to improve despeckling performance by alleviating local minima problem. All steps of the MND are parallelized using OpenMP on CPU and CUDA on GPU. Speckle reduction performance, execution time and memory consumption of the proposed method are shown using synthetic images and TerraSAR-X spot mode SAR images.

An implementation for quad-tree based solid object coloring using CUDA

Baha SEN, Caner OZCAN, Nesrin AYDIN ATASOY
Uluslararası BildiriProceedings on 2nd World Conference on Information Technology, Antalya, Turkey, 2011

Özet

We propose an implementation for quad-tree based solid object coloring using Compute Unified Device Architecture (CUDA). There are numerous different techniques in use for solid object coloring. One commonly used technique is the quad-tree, which has evolved from work in different fields. A quad-tree is a tree data structure in which each internal node has exactly four children. The quad-tree somewhat follows the tree data structure commonly used in computer science. The normal tree data structure looks like an upside down tree, where a parent node at the top of the tree has one or more children nodes connected to it. The aim of this study is coloring of a solid object using screen splitting method. The screen is divided into squares via this method and whether one or more points of the object are available in the separated parts is searched. According to the existing points, algorithm is applied and the object coloring is provided by reducing pixel size. We implemented our algorithm using the Graphics Processing Unit (GPU) computing and compared their performance with a CPU implementation. Nvidia CUDA library has been used for the GPU computing. CUDA gives developers access to the virtual instruction set and memory of the parallel computational elements in CUDA GPUs. We have tried our study on different systems that have different GPUs and CPUs. The computation studies were also evaluated for different solid objects. When we compared the results obtained from both systems, a better performance was obtained with GPU computing. According to results, GPU computation approximately worked 20 times faster than the CPU computation.

Performance comparison of gauss-jordan elimination method using OPENMP and CUDA

Baha SEN, Nesrin AYDIN ATASOY, Caner OZCAN
Uluslararası BildiriProceedings on 2nd World Conference on Information Technology, Antalya, Turkey, 2011

Özet

It is important to obtain the results of methods that are used in solving scientific and engineering problems rapidly for users and application developers. Parallel programming techniques have been developed alongside serial programming because the importance of performance has been increasing day by day while developing computer applications.Various methods such as Gauss Elimination (GE) Method, Gauss-Jordan Elimination (GJE) Method, Thomas Method, etc. have been used in solution of Linear Equation System (LES). In this study, performance comparison is done using Open Multi-Processing (OpenMP) and Compute Unified Device Architecture (CUDA) for nxn matrix via GJE Method. GJE Method is a variant of GE which is used in solving linear system equations (Ax=B). Each step of GJE Method solution algorithm is independent from each other and also the method is appropriate for parallel computing structure; therefore, this method is preferred within the scope of this study. Application coded in C programming language is developed using OpenMP and CUDA. OpenMP is an Application Program Interface that allows parallel programming using compiler directives on Central Processing Unit (CPU). CUDA is known as NVIDIA’s parallel computing architecture add it enables significant increases in computing performance by the Graphics Processing Unit (GPU).Application is realized on Intel Core 2 Quad CPU Q8200 2.33 GHz processor, GeForce 9500 GT graphic card. It is observed that application using Grid-Block-Thread structure and optimized with CUDA displays higher performance than OpenMP in terms of time.

Investigation of the performance of LU decomposition method using CUDA

Caner Ozcan, Baha Sen
Uluslararası BildiriProceedings on 2nd World Conference on Innovation and Software Development, Istanbul, Turkey, 2011

Özet

In recent years, parallel processing has been widely used in the computer industry. Software developers, have to deal with parallel computing platforms and technologies to provide novel and rich experiences. We present a novel algorithm to solve dense linear systems using Compute Unified Device Architecture (CUDA). High-level linear algebra operations require intensive computation. In this study Graphics Processing Units (GPU) accelerated implementation of LU linear algebra routine is implemented. LU decomposition is a decomposition of the form A=LU where A is a square matrix. The main idea of the LU decomposition is to record the steps used in Gaussian elimination on A in the places where the zero is produced. L and U are lower and upper triangular matrices respectively. This means that L has only zeros above the diagonal and U has only zeros below the diagonal. We have worked to increase performance with proper data representation and reducing row operations on GPU. Because of the high arithmetic throughput of GPUs, initial results from experiments promised a bright future for GPU computing. It has been shown useful for scientific computations. GPUs have high memory bandwidth and more floating point units as compared to the CPU. We have tried our study on different systems that have different GPUs and CPUs. The computation studies were also evaluated for different linear systems. When we compared the results obtained from both systems, a better performance was obtained with GPU computing. According to results, GPU computation approximately worked 3 times faster than the CPU computation. Our implementation provides significant performance improvement so we can easily use it to solve dense linear system.

Fuzzy Logic Graphical User Interface Software for Evaluation of Wooden Material Combustion Performance

Caner OZCAN, Raif BAYIR, Cemal OZCAN
Uluslararası BildiriProceedings on 2nd International Symposium on Computing in Science & Engineering, p:483-485, Kusadası, Turkey, 2011

Özet

In this study, fuzzy logic graphical user interface software for evaluation of wooden material combustion performance is realized. The most accurate monitoring of measurements obtained by the combustion is provided. Operations on parameters that obtained from the result of combustion can be made faster. Test performance is determined with fuzzy logic by using data obtained during the combustion process and errors are reported to the user. Unexpected situations that experts could not notice can be determined easily by this software during the combustion. It has been observed that this user interface software prevents the data-loss and gives better results with sensitive measurements. Since repetition of the experiment is reduced, especially time, work and energy savings are provided. This fuzzy logic user interface software with these features can be used easily in wooden material combustion R&D centers, academic studies on this issue and companies in this area.

Real Time Control via Matlab/Simulink of DC-DC Buck Converter with Fuzzy Logic Controller

Caner Ozcan, Ali Uysal, Raif Bayir
Uluslararası BildiriFUZZYSS'09: 1st International Fuzzy Systems Symposium, Ankara, Turkey, 1-2 October, 2009

Özet

In this study, the real time control of DC-DC buck converter with fuzzy logic controller is presented. DC-DC converter circuit and real time Matlab/Simulink model that works with this circuit is prepared. For real time control, it is benefited from Real Time Windows Target library. Data acquisition card is used for data transfer between Simulink model and designed circuit. Fuzzy inference system prepared with FIS editor is embedded to fuzzy logic controller used in this model. The real time and non-real time simulations of fuzzy logic controller and PID comparisons are given with this study.

Spark Destek Vektör Makinesi ile Metin Sınıflandırma

İskender Ülgen Oğul, Caner Özcan, Özlem Hakdağlı
Ulusal Bildiri1.ULUSAL BULUT BİLİŞİM VE BÜYÜK VERİ SEMPOZYUMU B3S'17, Antalya, Türkiye, 19-20 Ekim, 2017

Özet

With the accumulation of technology from the recent years of development, the total amount of data that has been revealed daily at the point reached today are quite excessive in the last forty years. It is of vital importance in our day to clean, understand, classify and make a usable analysis report of this emerging data. Speed is a key factor for information retrieval in this important area. Apache Spark offers an innovative and robust solution to this problem. Unlike the previously used analysis methods, this solution stores data in the memory in order to gain speed in analysis, rather than performing data read/write operations through the disk during analysis. Spark provides near real-time speed with in-memory data read/write, analysis and resulting operations. In this study, a classification analysis was conducted on Reddit social entertainment site comments using Support Vector Machine (SVM) which is the supervised (educated) learning method. The results are presented in tables and graphics.

Early-exit Optimization Using Mixed Norm Despeckling for SAR Images

Ozcan, Caner; Sen, Baha; Nar, Fatih
Ulusal BildiriIEEE Signal Processing and Communications Applications Conference (SIU), 2015 23nd, Page(s): 685 - 688, Malatya, Turkey, 16-19 May, 2015

Özet

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 sensing applications, efficiency of computational load and memory consumption of despeckling must be improved for SAR images. In this paper, an early-exit total variation approach is proposed and this approach combines the l1-norm and the l2-norm in order to improve despeckling quality while keeping execution times of algorithm reasonably short. Speckle reduction performance, execution time and memory consumption are shown using spot mode SAR images.

Fast feature preserving despeckling

Ozcan, Caner; Sen, Baha; Nar, Fatih
Ulusal BildiriIEEE Signal Processing and Communications Applications Conference (SIU), 2014 22nd, Page(s): 1007 - 1010, Trabzon, Turkey, 23-25 April, 2014

Özet

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 is quite important. Besides, due to huge size of SAR images in remote sensing applications efficiency of computational load and memory consumption must be further improved. In this paper, a parallel computational approach is proposed for the Feature Preserving Despeckling (FPD) method which is chosen due to its success in speckle reduction. Speckle reduction performance, execution time and memory consumption of the proposed Fast FPD (FFPD) method is shown using spot mode SAR images.

Kümelenmiş sanal sınıf uygulaması

Şen, B., Menemencioğlu, O., Atasoy, F. ve Özcan, C.
Ulusal BildiriAkademik Bilişim 2011, Malatya, 1-8 (2011)

Özet

Açık kaynak kodlu işletim sistemi, öğrenim yönetim sistemi ve sanal sınıf uygulaması ile gerçekleştirilen makul sayıda öğrencinin olduğu bir uzaktan eğitim altyapısında: hizmet kalitesini arttırmanın ve hizmet verilen öğrenci sayısını arttırarak bunu gerçekleştirmenin yolları irdelenmiştir. Çalışma özellikle sanal sınıf uygulamalarında performans artırımı ile ilgilenmektedir. Yeni, daha güçlü sunucular alarak başarım sağlamak her kurumsal yapıya uydurulamayacağından veya her kurum vizyonu yenilendiğinde ona uygun alt yapı değişikliği gerektirdiğinden, aşağıda da detaylandırıldığı üzere buna alternatif yöntem önerilmiş, modellenmiş ve gerçekleştirilmiştir.

Karabük Üniversitesi Mühendislik Fakültesi Bilgisayar Mühendisliği Bölümü Lisans/Y.Lisans/Doktora Programlarının Tanıtımı

Baha ŞEN, Caner ÖZCAN
Ulusal BildiriElektrik Elektronik Bilgisayar Biyomedikal Mühendislikleri Eğitimi 4. Ulusal Sempozyumu, Eskişehir, 22-24 Ekim 2009

Özet

Bilgisayar Mühendisliği, bilgisayar ve ilgili cihazların tasarımı, üretimi, işletimi ve bakımı ile ilgili mühendislik dalıdır. Donanım, yazılım, iletişim ve bunların arasındaki etkileşimleri inceler. Bilgisayar mühendisliği öğrencileri, teoriler, ilkeler, geleneksel elektrik mühendisliği pratikleri ve matematiğe odaklanır ve bunları bilgisayarlar ya da bilgisayar temelli sistemler tasarlama problemlerine uygular. Eğitimin ilk yıllarında matematik, fizik, kimya ve bilgisayar mühendisliğine giriş dersleri, daha sonraki yıllarda sistem programcılığı ve donanım kısmını oluşturan ve uzmanlaşmayı gerektiren dersler verilmektedir. Bilgisayar mühendisliğinde, programlama dilleri, veri yapıları, olasılık ve istatistik, sayısal mantık sistemleri, elektrik devreleri, temel elektronik, bilgisayar mimarisi, mikroişlemci ve mikro bilgisayar, sistem programlama, veri tabanı yönetim sistemi, bilgisayarlı grafik, işletim sistemi, yöneylem araştırma, formal diller ve soyut bilgiler, dil işleyiciler, bilgi sistemleri mühendisliği ile uygulamaya yönelik program yürütülmektedir. Lisans programımız, bilgisayar mühendisliğinde gerekli tüm temel bilgileri vermenin yanı sıra, profesyonel mühendislik için gerekli olan teknolojiyi yakından takip eden genç ve dinamik öğretim üyesi kadrosuyla kısa süre içerisinde gerçekleştirmeyi başarabilecek düzeydedir. Bölümümüze 2009-2010 eğitim-öğretim yılından itibaren normal öğretim ve ikinci öğretim olmak üzere 80 öğrenci alınacaktır. Bölümümüzde 2009-2010 yılından itibaren ülkemizde birkaç üniversitede uygulaması bulunan Uzaktan Eğitim Merkezi aracılığıyla uzaktan eğitimle Bilgisayar Mühendisliği Lisans programında bilgisayar mühendisi olma fırsatı sunulmaya başlanacaktır. 2008-2009 eğitim-öğretim yılından itibaren yüksek lisans ve doktora programları açılmış olup, mezun olacak öğrencilerimiz akademik çalışma yapabilme fırsatına sahiptirler.

Machine learning methods in image processing using Apache Spark cluster computing

Caner Ozcan, Okan Ersoy
Diğer BildiriPurdue University International Scholar Research Symposium, West Lafayette, IN, USA, March 14, 2018

Özet

Synthetic-aperture radar (SAR) image classification plays an important role in the analysis and interpretation of SAR images. In general, SAR image classification has a wide field of application in many areas. However, the analysis of high-resolution SAR images, such as classification, is a time-consuming process due to its size. In this study, SAR image patch classification is provided efficiently using Apache Spark framework. A successful classification was performed using the Naive Bayes method in the Spark MLlib library for image classification. Feature vectors required for each SAR image patch are obtained using gray-level cooccurrence matrix (GLCM).

CPU and GPU Parallelization of Sparsity Driven Despeckling in SAR Images

Ozcan, Caner; Sen, Baha; Nar, Fatih
Diğer BildiriThe International Remote Sensing Conference in Saudi Arabia , Riyadh, Saudi Arabia, 17-209 Jan, 2016

Özet

In this paper, we present the implementation details and CPU/GPU parallelization of Sparsity Driven Despeckling (SDD) method which we proposed before. In this method, a sparsity driven total variation (TV) approach employing l0-norm, fractional norm, or l1-norm in order to smooth homogeneous regions with minimal degradation in edges and point scatterers is proposed. An observation image corrupted by speckle noise is taken into consideration and SAR image despeckling problem is defined as the optimization problem with minimization of the proposed cost function. In order to use convex optimization methods, corresponding cost function is approximated and written in a matrix-vector form. This matrix-vector form enables a special iterative optimization method where a linear system is solved in each step. Linear system which consists of positive definite symmetric matrix is solved efficiently using preconditioned conjugate gradient (PCG) iterative solver. PCG with IC preconditioner is implemented single threaded since construction of IC preconditioner, lower and upper Gaussian eliminations are not efficiently parallelizable. PCG with Jacobi preconditioner is parallelized using OpenMP on CPU and CUDA on GPU since all employed operations are efficiently parallelizable. All step of the preconditioned conjugate gradient algorithm is also parallelized using OpenMP and CUDA. During the PCG algorithm, new strategies were developed to minimize the transfer from the CPU to the GPU. We used tiling approach for large images which PCG consumes too much memory. Therefore streams which have an ability to perform multiple CUDA operations simultaneously are used to run multiple tiles concurrently. Dynamic parallelism that enables a CUDA kernel to create and synchronize new nested work is designed. We realized our studies with two versions of the developed method using and not using dynamic parallelization. We performed our test studies with GeForce and Tesla GPUs and we also used Nvidia Jetson TK1 development kit for mobile platforms. In consequence, the results show that our method provides extremely fast (near-real-time) and high-quality SAR despeckling. All the alternative implementations of the SDD method are tested and information is given about cases in which ones could be chosen. Despeckling performance, execution time and memory consumption of the proposed method are shown with detailed tables and figures using synthetic images and real-world SAR images.

Dönem Dersleri

  • Y.Lisans 2018/19 Güz

    BSM754-Sayısal Görüntü İşleme Uygulamaları

    Bu dersin amacı, görüntü işlemede kullanılan temel ve ileri düzey işlemleri öğrenmek ve uygulayabilmek. Öğrencilerin, bu ders sayesinde görüntü analizi işlemlerini gerçekleştirebilmeleri ve elde edilen sonuçları tartışabilmeleri amaçlanmaktadır.

 

  • Lisans 2018/19 Güz

    BLM429-Sayısal Görüntü İşleme

    Bu dersin amacı görüntü işlemede kullanılan temel prensipler ve algoritmaları öğrencilere öğretmektir. Öğrencilerin, bu ders sayesinde görüntü analizi işlemlerini gerçekleştirebilmeleri ve elde edilen sonuçları tartışabilmeleri amaçlanmaktadır.

    ~~SUNUM BİLGİLERİ (PRESENTATION INFORMATION)~~

    KONULAR (TOPICS) - ÖĞRENCİ LİSTESİ (STUDENT LIST)

    TR Sunum Arşivi (TR Presentation Archive) - ING Sunum Arşivi (ENG Presentation Archive)

    ÖDEVLER:

    Ödev 1 / Ödev 2

    Ders Sunumları:

    Hafta 1. Sayısal Görüntü İşlemeye Giriş: İngilizce - Türkçe

    Hafta 2. Görüntünün Alınması ve Sayısallaştırılması: İngilizce - Türkçe

    Hafta 3. Görüntü İşleme ile İlgili Temel Kavramlar: İngilizce - Türkçe

    Hafta 4. Yoğunluk Dönüşümleri ve Histogram İşlemer: İngilizce - Türkçe

    Hafta 5. Uzamsal Filtreleme: İngilizce - Türkçe

    Hafta 6. Görüntü Onarma ve Geriçatma: İngilizce - Türkçe

    Hafta 7. Renkli Görüntü İşleme

    Hafta 9. Morfolojik Görüntü İşleme

    Hafta 10. Görüntü Sıkıştırma

    Hafta 11. Görüntü Bölütleme

    Ek Kaynaklar:

    Image Processing Toolbox User's Guide: Link

    An Introduction to Digital Image Processing with Matlab: Link

    Intro to Digital Image Processing: Online Course

    Image Processing Using Matlab: Online Course

    E-book: The pocket handbook of image processing algorithms: Link

    E-book: Image Processing in C: Link

    Opencv C++ Eğitimi: Link

    OpenCV 2.4 Cheat Sheet (C++): Link

    E-book: Learning OpenCV: Link

    Image and video processing: From Mars to Hollywood: Video Link

    Dijital Görüntü ve Video İşlemenin Temelleri: Online Course

  • Y.Lisans 2016/17 Bah.

    BSM744-Uzaktan Algılama ve Uygulamaları

    Bu dersin amacı, uzaktan algılamanın temel kavramları ile uydu görüntülerinin işlenmesi, yorumlanması ve analizi işlemlerini öğrencilere öğretmektir.

  • Lisans 2016/17 Bah.

    MSD302 Araştırma ve Sunum Teknikleri

    Bu dersin amacı, öğrencilere bilimsel araştırma ve inceleme tekniklerinin öğretilmesi, elde ettikleri verilerin kullanılmasını ve sunulmasının öğretilmesi.

  • Y.Lisans 2016/17 Güz

    BSM754-Sayısal Görüntü İşleme Uygulamaları

    Bu dersin amacı, görüntü işlemede kullanılan temel ve ileri düzey işlemleri öğrenmek ve uygulayabilmek. Öğrencilerin, bu ders sayesinde görüntü analizi işlemlerini gerçekleştirebilmeleri ve elde edilen sonuçları tartışabilmeleri amaçlanmaktadır.

  • Lisans 2016/17 Güz

    BLM111-Programlama Dilleri I

    Bu dersin amacı, problem çözümüne yönelik algoritma tasarlama ve C programlama dili kullanarak program geliştirmenin öğrenilmesi. C programlama tekniklerini kullanarak çeşitli problemlere çözüm bulunması.

  • Lisans 2016/17 Güz

    BLM429-Sayısal Görüntü İşleme

    Bu dersin amacı görüntü işlemede kullanılan temel prensipler ve algoritmaları öğrencilere öğretmektir. Öğrencilerin, bu ders sayesinde görüntü analizi işlemlerini gerçekleştirebilmeleri ve elde edilen sonuçları tartışabilmeleri amaçlanmaktadır.

  • Y.Lisans 2015/16 Bah.

    BSM744-Uzaktan Algılama ve Uygulamaları

    Bu dersin amacı, uzaktan algılamanın temel kavramları ile uydu görüntülerinin işlenmesi, yorumlanması ve analizi işlemlerini öğrencilere öğretmektir.

  • Lisans 2015/16 Bah.

    BLM112-Programlama Dilleri II

    Bu dersin amacı, işaretçiler ve listelerin kullanımını anlamak. Basit sıralama ve arama algoritmalarını öğrenmek. Öğrencilerin dosya işlemleri, bitişlemler, görsel programlama ve temel grafik işlemleriyle etkin programlar yazmasını sağlamaktır.

Geçmişte Asistanlık Yapılan Dersler

  • 2015 2013

    Database Management

    The purpose of this course is to get students to gain the ability of modelling standard and data-based technology. Turning the database of Entity-Relation or Object-Role into relational modelling.

  • 2015 2013

    Data Structures

    The purpose of this course is to teach data structures with C programming language in details.

  • 2015 2012

    File Organization

    Introducing the most important high-level file structures tools which include indexing, co-sequential processing, B trees, Hashing.

  • 2012 2010

    Veri Yapıları ve Algoritmalar

    Dersin amacı gerçek hayat problemlerinden hareketle, farklı alanlarda kullanılabilecek algoritmaların sunulmasıdır.

  • 2010 2009

    Programlama Dilleri II

    Bu dersin amacı öğrencilere C programlama dilinde işaretçi (pointer) kullanımı, bit düzeyinde işlem yapma, dosya işlemleri, sıralama algoritmaları, görsel programlama konularında yetenek kazandırmaktır.

  • 2010 2009

    Programlama Dilleri I

    Bu ders öğrencilere C programlama dilini kullanarak yapısal programlamanın temel kavramlarını öğretir.

Ofis Çalışma Saatleri