Bilgisayar Mühendisliği Bölümü Tezleri

Permanent URI for this collectionhttps://hdl.handle.net/20.500.12416/58

Browse

Now showing 1 - 20 of 270

Attention Mekanizmaları ve Hibrit ViT-ResNet Mimarisi ile Gemi Görüntülerinin Çok Sınıflı Sınıflandırılması
(2025) Ergün, Berkay; Arslan, Serdar
Bu tezde, gemi görüntülerinin çok sınıflı sınıflandırılması için Vision Transformer (ViT) ve ResNetRS50 tabanlı hibrit bir model geliştirilmiştir. ViT yüksek seviyeli anlamsal bilgileri, ResNetRS50 ise düşük ve orta seviyeli mekânsal özellikleri çıkarmakta; bu iki yapı, dikkat (attention) mekanizmaları ve Gated Fusion katmanı ile birleştirilmektedir. Eğitim sürecinde MixUp ve CutMix veri artırma yöntemleri, Focal Loss ile bilgi aktarımı (distillation) kaybı, OneCycleLR zamanlayıcı, otomatik karma hassasiyet (AMP) ve model ağırlıklarının üssel hareketli ortalaması (EMA) kullanılmıştır. Sekiz gemi sınıfından oluşan veri kümesi üzerinde yapılan deneyler, önerilen mimarinin hem doğruluk hem F1 skoru açısından tek başlı CNN veya ViT modellerinden daha yüksek performans gösterdiğini ortaya koymuştur. Sonuçlar, hibrit mimariler ve dikkat tabanlı füzyon stratejilerinin gemi sınıflandırma problemlerinde etkin bir çözüm sunduğunu göstermektedir.
Integrating computer vision with a robot arm system
(2014) Yosif, Zead Mohammed
During last decades, robotic system has been employed in different fields, such as, industrial, civil, military, medical, and many other applications. Vision system is integrated with robot systems to enhance the controlling performance of the robot system. A great deal of features can be computed using the information have been gotten from vision sensors (camera). The extracted information from vision system can be used in the feedback to have the ability to control the robot armtor motion, but the operation of extracting this information from vision system is time consuming. This thesis addressed the problem of following (tracking) and grasping of moving target (object) with limited velocity in real time by employing the technology of Eyein- Hand, whereas a camera attached (mounted) to the robot arm end effector. This done by using a predictor (Kalman filter) that estimates the positions of the target in the future, an algorithm was designed to track an object move in different trajectories, within the camera field of view. The Kalman filter uses the measured position of the target as well as previous state estimates to fix the location of thetarget object at the next time step, in other word, the Kalman filter is applied to keep observing the object till grasp it. The employing of vision system information in the feedback control of the robot systems have been the major research in robotics and Mechatronic systems. The utilizing from this information has been proposed to handle stability and reliability issues in vision-based control system.
Evaluation of terrain rendering algorithms
(2005) İnam, Emin
errain rendering plays an important role in outdoor virtual reality applications, games, Geographic Information System (GIS), military mission planning's and flight simulations, etc. Many of these applications require real-time dynamic interaction from end users and thus are required to rapidly process terrain data to adapt to user input. Typical height fields consist of a large number of polygons, so that even most high performance graphics computers have great difficulties to display even moderately sized height fields at interactive frame rates. The common solution is to reduce the complexity of the scene while maintaining a high image quality. This thesis is an evaluation of three real-time continuous terrain levels of detail algorithms described in the papers ROAMing Terrain: Real-time Optimally Adapting Meshes by Duchaineau, Real-Time Generation of Continuous Levels of Detail for Height Fields by Röttger and Fast Terrain Rendering Using Geometrical MipMapping by Willem H. de Boer. The evaluation and comparison of the algorithms is based on the trade- off of polygon count to terrain accuracy over separate test data sets. The main aim of this thesis is research on terrain rendering algorithms that is generate high quality image in real-time with using height data.
Multifunction robot controlled by computer vision system
(2014) Mustafa, Mohammed Sulaiman
In this thesis, we try to come up and build a robot platform with multifunction capabilities, easy to add, modify and delete those functions without redesigning, by using easy use technology that can create a suitable efficient platform. The process of building platform is by using figures, tables and programming code to make this thesis capable to apply and implement in real world, showing obstacles and challenges that lead to the key of success until it reaches the final goal. This thesis requires only basic level in electronic and computer programming because we are using a simplified way for building robot. The multifunction platform is a unique idea and opens new space to experimenters to get benefits from this opinions or ideas to use these functions in raw state, with no need to study hardware and software material of robot. The final robot form is shown in the last pages of this thesis as appendix
Classification of diabetic retinopathy using pre-trained deep learning models
(2019) Al-Kamachy, Inas Mudheher Raghib Kafı
Diabetic Retinopathy (DR) is considered to be the first factor that leads to blindness. If it is not detected early, many people around the world would suffer from the diabetic disease that may lead to DR in their eyes. Any delay in regular monitoring and screening by ophthalmologists may cause rapid and dangerous progress of this disease which finally leads to human vision loss. The imbalance between the numbers of doctors required to monitor this disease and the number of patients around the world increasing year by year shows a major problem leading to poor regular monitoring and loss vision in many cases which could have been detected had there been good treatment in the earlier stages of DR. In order to solve this problem, serious aid was needed for a computer aid diagnosis (CAD). Deep learning pre-trained models are state-of-art in image recognition and image detection with good performance. In this research, we used image pre-processing and we built several convolution neural network models from scratch and fine-tuned five pre-trained deep learning models which used ImageNet as the dataset for medical images of diabetic retinopathy in order to classify diabetic retinopathy into five classes. After that, we selected the model that showed good performance to build a diabetic retinopathy web application using Flask as a framework web service. We used the KAGGLE kernel website with Jupyter as a notebook as well as Flask to build our web application. The final result of the AUC was 0.68 using InceptionResNetV2.
Web services based real time data warehouse
(2012) Obalı, Murat
Today's business environment is quickly changing and business decision makers need for a historical picture of what happened and a picture of what was happening today. Traditional data warehouses provide a historical picture, but there is lack of fresh data. However, fresh data in data warehouses is a strong feature from the part of the users. The aim of this study is building a real time data warehouse using web services. First, we modelled both the conceptual and the logical design of real time data warehouse. For change data capture from source systems, we implemented web services based server and client software. Then, we used real time partition for real time data which is merged into data warehouse in a daily fashion. We, also, implemented a data integration service using query re-write approach to integrate data warehouse and real time partition data.
Parallelization study on the clustering technique to mine large datasets
(2011) Yıldırım, Ahmet Artu
Parallel clustering algorithm implementations concerning message passing interface (MPI) and compute unified device architecture (CUDA) model with their applications to very large datasets have been presented in the thesis. WaveCluster is a novel clustering approach based on wavelet transforms. Despite it?s novelty, it requires considerable amount of time to collect results for large sizes of multidimensional datasets. In the MPI algorithm; divide and conquer approach has been followed and communication among processors are kept at minimum to achieve high efficiency. Developed parallel WaveCluster algorithm exposes high speedup and scales linearly with the increasing number of processors. Parallel behavior of WaveCluster approach has been also investigated by executing the algorithm on graphical processing unit (GPU). High speedup values have been obtained in the computation of wavelet transform and connected component labeling algorithms in the GPUs with respect to the sequential algorithms running on the CPU
Model based human face detection using skin color segmentation
(2002) Özbay, Eylem
Model Based Human Face Detection Using Skin Color Segmentation Özbay, Eylem Ms, Department of Computer Engineering Supervisor: Dr. Reza Hassanpour January 2005, 85 pages For identification of the people easiest way using the faces. However, it requires determining the location of the faces in the images. Face Identification systems are generally preceded by face segmentation systems. The main goal of thesis is locating the human faces and segmentation regions belonging to them using skin color segmentation methods and facial features such as nose, eyes, mouth etc. The segmentation results may be used as input to other related systems
Derin Öğrenme ve Çok Boyutlu İndeksleme Kullanılarak İçerik Tabanlı Görüntü Alma
(2024) Uzel, Ömer; Arslan, Serdar
Son yıllarda yaşanan teknolojik gelişmeler ile donanım ve yazılım maliyetlerindeki düşüş, görsel arama uygulamalarını hem popüler hem de vazgeçilmez bir hale getirdi. Dolayısıyla, görsel sorgular aracılığıyla görüntülerin geniş veri tabanlarından hızlı ve hassas bir şekilde alınması kritik bir görev haline geldi. Video karesi düzeyinde veri tabanı aramaları yürütülen sistemlerle karşılaştırıldığında, arama performansını önemli ölçüde artıran yeni bir system sunuyoruz. Önceden eğitilmiş bir Evrişimli Sinir Ağı (CNN) modelinden yararlanarak, verimli indeksleme için düşük seviyeli özellikleri çıkarmak ve depolamak amacıyla denetimsiz görüntü alma süreçlerini kullanıyoruz. Hızlı ve etkili erişimi kolaylaştırmak için, Bakış Noktası Ağacı (VP Tree) olarak bilinen düşük seviyeli özelliklerden yararlanan bir indeksleme yapısı uyguluyoruz. Bu özelliklerden faydalanabilmek için, onları daha düşük boyutlu bir alanda temsil edecek boyut küçültme tekniklerini kullanıyoruz. Karşılaştırmalı görüntü veri kümesi üzerinde gerçekleştirilen deneylerimiz, bu yaklaşımın, K-En Yakın Komşu (KNN) araması olarak bilinen bir arama yöntemiyle karşılaştırıldığında daha hızlı ve doğru erişime yol açtığını göstermektedir. Ayrıca, önerilen tekniği iki gerçek video veri kümesini kullanarak KNN'e karşı değerlendiriyoruz ve bu teknik, sürekli olarak KNN'den daha iyi performans gösteriyor.
Design and analysis of native XML databases in three-tier architectures
(2004) Ergen, Mehmet Tunç
XML is rapidly emerging as a standard for exchanging business data on the World Wide Web. From management systems to e-business application providers to pure development tools, XML has gone from newly underground technology to integrated component standard. It is used as the file format of choice for Web development, document interchange, and date interchange, and presents a new world of opportunities and challenges to programmers. It is predicted that by at the end of 2004, more than 75% of e-business applications will include XML regardless of which language the application has been written in. As more and more applications starts using XML there wilt be a need to efficiently handle the XML data at the back-end. The need to efficiently store and process XML documents has created the new XML supported technologies and tools. One of these tools is the Native XML Databases. It is based on document-in, document-out architecture with capabilities for storage, retrieval, querying and updating the documents. While Native XML Databases are an important new technology, they should not be used without careful analysis and consideration. In this thesis Native XML Databases are investigated and analyzed in a 3-tier architecture to gain and ensure several advantages that three-tier systems offer to application developers and information technology industry
Comparative study of human face identification methods
(2005) Karaömeroğlu, Betül
Face recognition algorithms have gained popularity in the recent years. Nowa days there are a lot of face recognition systems. However, it is known that these systems work properly in ideal environments. Behind of that, these applications provide low performance in some situations such as illumination, occlusion, different facial expression or presence of rotation and tilt conditions. In this thesis, a new approach to face recognition problem has been proposed. The presented system is a specialized version of PCA augmented with Gabor Wavelet Transform. Firstly, 2D Gabor Wavelet Transform is applied to cope with the variations due to illumination and facial expression changes; then the modified PCA approach is used for reducing a large set of correlated variables iii to a small number of uncorrelated components. The performance of the proposed algorithm is compared with the other algo rithms based on the effect of the illumination, facial expression and occluding objects such as eye glasses and facial hair
Performance comparison of routing algorithms used in packet switching computer networks
(2015) Yaseen, Nawfal
Bu tez ağ üzerindeki trafik mühendisliğinin daha iyi sağlanabilmesi için kullanılan Çok Protokollü Etiket Anahtarlama (MPLS) mimarisi tabanlı rotalama algoritmalarının incelenmesini ve karşılaştırılmasını içermektedir. Ağlar üzerinde yapılan bu çalışmada sürekli olarak büyüyen internet servislerinin Trafik Mühendislği üzerine etkilerine odaklanılmaktadır. MPLS, ağ üzerinde kullanılan yönlendiricilerin hızını arttıracak ve gerçek zamanlı çoklu ortam verilerinin aktarılacak şekilde ağ servislerini sağlayan modern bir teknik kullanır. Ayrıca, MPLS mimarisi ağ üzerinde bulunan farklı yönlendiricileri tanımlar ve trafik yönetiminin sağlanmasını garanti altına alır. MPLS'in günümüzde en kısa yoldan gönderim yapılacak yulu seçen birçok protokol içinden en iyier inden birisidir. Bir diğer önemli nokta ise, İnternet Servis Sağlayıcılarının (ISP) MPLS yönlendiricileri üzerinde çalıştırılan rotalama algoritmalarını geliştirme gereksinimlerinin olmasıdır. Bu algoritma sade ve yönlendiricilerde uygulanabilecek nitelikte olmalıdır. Gelişmiş rotalama algoritmaları, MPLS ağının rotalama algoritmalarını geliştirme avantajlarını da kullanırlar. MPLS üzerinde araştırılmaya devam edilen ve kuuanılan birçok gelişmiş rotalama algoritması bulunmaktadır. Bunlardan en önemlileri ise, trafik durumuna göre karar vererek bloke olma olasılığını azaltan rotalama teknikleri ile her türlü etiket anahtarlamalı yol (LSP) isteği için çözüm sunan trafik tabanlı algoritmalardır. Bu nedenle bu tezde, bilgisayar ağlarında iyi bilinen bir protokol MATLAB ile benzetim yoluyla çalışılmış ve daha sonra bazı farklı rotalama teknikleri bu protokol üzerinde uygulanmıştır. Son olarak, simulasyon sonuçları incelenmiş ve farklı bakış açılarından analiz edilmiştir. Çalışmanın sonunda, farklı ağ koşulları için; incelenen rotalama algoritmalarından bazıları önerilmiştir. Bu benzetim sonuçları, paket anahtarlamalı bilgisayar ağlarında entegre edilerek kullanılan rotalama algortimalarının avantajlarını göstererek ortaya koymaktadır.
Application layer processing with protocol independent switch architecture
(2021) Tuncel, Yusuf Kürşat
Bu tez, uygulama katmanı verilerini işlemek ve bir uygulama içeriğinin incelenmesini sağlamak için Protokolden Bağımsız Anahtar Mimarisi için bir çözüm araştırır ve bir yöntem önerir. Protokolden Bağımsız Anahtar Mimarisi, ağ anahtarının herhangi bir gömülü ikili kod çalıştırmadığı, bunun yerine amaca özel bir dilde yazılmış yorumlanmış bir kod çalıştırdığı ağ iletişiminde yeni bir yaklaşımdır. Bu yaklaşımın arkasındaki ana motivasyon, telekomünikasyon operatörlerinin herhangi bir ağ ekipmanı türü için bir satıcıya kilitlenmek istememeleri, tek bir ekipman üreticisi tarafından yönetilmeyen bir donanım ekosisteminde kendi ağ kodlarını geliştirmeleridir. Bu yaklaşım aynı zamanda, bir donanım anahtarının tüm bileşenleri yazılımla modellenen bir anahtarda aynı uyumlu kodu çalıştırdığı için bir ekipmanın simülasyon ortamında modellemesini kolaylaştırır. Bu tezdeki yeni teknikler, programlanabilir bir anahtarın ana işlevlerinden yararlanarak, maliyetleri düşürmek ve ağı kapsamlı bir şekilde yönetmek için bir telekomünikasyon operatörü perspektifinden istenen etkiyi yaratarak akışlı veri işlemci yazılımını ağ anahtarı yazılımıyla birleştirmeyi amaçlamaktadır. Deneysel çalışma, herhangi bir optimizasyon yapılmadan önerilen çözümün uygulama tanımlama sistemlerinin performansını 5.5'ten 47 katına çıkardığını göstermektedir.
Defect product estimation using customer reviews, Amazon use case
(2022) Eyerci, Tarkan
Teknoloji her alanı etkilediği gibi ticareti de çok etkiledi. Günümüzde artık, üreticiler, perakendeciler, hizmet sağlayıcılar gibi son kullanıcıya hitap eden tüm işletmeler e-ticaret siteleri ve mobil uygulamaları gibi yöntemlerle internet üzerinden müşterilerine hızlıca ulaşabiliyorlar. Diğer yandan, müşteriler ise artık birçok seçenek arasından seçim yapma şansına sahipler. Kullanıcılar genellikle seçimlerini yaparken daha önce aynı tecrübeyi paylaşmış diğer kullanıcıların yorumlarından faydalanırlar. Bu açıdan kullanıcı yorumları çok değerli bilgiler içerir. Fakat yoğun kullanılan sitelerde bir insanın tek tek inceleyemeyeceği kadar çok yorum birikir. Biz bu çalışmada, ürünlerin belli bir özelliğine, yani kusurlu özelliklerine odaklandık. Kusur bilgisi içeren milyonlarca yorum içinden ilgili yorumları filtre edebilmek için bir yöntem öneriyoruz. Kusur ile ilgili kelimeleri sözlük yardımı ile elle oluşturup bu kelimeler geçen yorumları filtrelemek bir çözüm önerisi olabilir. Fakat bu kelime listesini elle oluşturmak verimli olmayacaktır. Bunun için sadece ilgili ürün gruplarına ait yorumları kullanarak kendi kelime temsil modelimizi eğitip, bu modelle birlikte kelime yakınlıklarını kullanarak daha verimli bir kusur kelimeleri listesi oluşturduk. Kullanıma hazır önceden eğitilmiş bir kelime temsil modelini indirip, bu modelle kendi modelimizi kıyasladık. Genel konularda hazır modelin daha başarılı olurken, özel bir konuda kendi modelimizin kelime listesi oluşturmada daha başarılı olduğunu gördük.
Smart home design with IoT (Home automation)
(2018) Hadi Al-Samarraei, Yasser Samir
Smart home's conception has been growing topic in our modern community. The content of smart automated items, such as home electronic devices, communicate together with users. These devices are quite expensive when compared to non-intelligent devices. In this thesis, a Home Automation system has been studied and developed to automate existing appliances by supporting them with modern smart technology. In this effective way, the cost can be lowered with upgrading the automating system instead of replacing entire the items. A Home Automation System has been implemented and designed to solve several common issues in this thesis. IoT security items have been investigated with the smart house (Home Automation) and clarified to become a basic stage of the Home Automation system. Smart home (Home Automation) application is designed to monitor and secure the home security sensors. An application has been performed to monitor and control the smart home system. The system is based on hardware and software components (such as Arduino Mega 2560) and can be acted as a security guard of the home. The system has been designed to monitor the humidity, temperature, light, gas, fire, burglar alarm in the house and has a passive-infrared sensor to make sure that the family in safe. Sensors have used on windows and doors to monitoring any an unauthorized intruder or in case of forgetting to close it. In addition, the system is also connected with GSM system (Global System for Mobile communication) to send notification SMS message (Short Message Service) to specific destinations in case of detecting something from sensors while "the system activated" mode on. The system (equipment of the home design) also has been controlled by using android application via Bluetooth
Matching composite drawings and mugshot photographs to determine the identity of the person
(2019) Karasolak, Mustafa
In this thesis, a new photo-sketch generation and recognition technique is proposed using residual convolutional neural network architecture. For this, the proposed architecture is trained with face photos and sketches. Sketches are applied to the proposed Region-based Convolutional Neural Networks (RCNN) architecture and, face photos are obtained at network output. Then, the obtained face photographs are compared with the images in the database. It is associated with the highest similarity photograph. Structural Similarity Index (SSIM) is used to measure similarity. It is very useful for law enforcement for image processing applications. 188 images are used for training and testing. Of these, 148 are used for training. 20 are used for validation and 20 are used for testing. Data augmentation is applied to 148 images used for training. As a result of the data augmentation process, 444 face images are obtained and used for network training. As a result of network training, the success of the training curve is 90.55% and the validation success is 91.1%. True face recognition success from generated face images with SSIM is 93.89% for CUHK database and 84.55% AR database.
IP traffic modeling
(2007) Tarıyan, Sibel
RSVP allows Internet real-time applications to request a specific end-to-end QoS for data stream before they start transmitting data. In this report firstly an overview of RSVP is presented. After that the different quality of services available and the relation between QoS and RSVP are explained. Then the fundamentals of RSVP as a protocol are discussed. The performance issues and benchmarking are given next. The experimental results and discussions conclude this thesis
Decision making via the analysis of school locations in Kirkuk province using gis technology
(2015) Hashim Shaker, Ehab
GIS technology is increasing in importance and now affects many aspects of daily life. In recent decades, GIS technology has entered many service fields, in particular health and education, two of the most popular services in any community. GIS technology is often employed to both regulate these services and to make better decisions. Furthermore, in order to construct a new and modern city in any country, GIS technology is typically used to achieve homogeneity of service distribution for the benefit for all citizens across society. The present thesis thus examines the challenges currently facing the city of Kirkuk in terms of service distribution, especially regarding education services. The results show that schools in the study area suffer from poor distribution, putting pressure on the service providers and adversely affecting the standard of education. Recommendations based on this research include the need to develop a plan aimed at the development of education services which is in line with both planning and exceptional standards. Work should be undertaken to establish new school sites in order to meet the shortfall currently experienced in the study area, taking into account urban development and future population growth. Finally, this thesis identifies those schools which are either directly or indirectly at risk due to their proximity to institutions targeted by terrorist operations.
Moving object detection in industrial line application
(2014) Abdilatef, Muhamad Azhar
In this thesis, a comparison study of moving object detection methods in industrial line application is presented. This comparison includes the consuming times, and the detection accuracy. According to consuming time, the methods are sorted into three groups (A, B, C). Groups A and B are including the methods those consuming a large time, thus they didn’t used for our application, while group C includes the methods with low consuming time. According to the detection accuracy, group C methods are compared one with each other to select the best method. The Misclassification rate (MR) is used to do this comparison. Some applied methods gave good results in detection but with high consuming time, others have problems in detection but with low consuming time. In this thesis a new method is presented using a combination between method 5 (the statistical morphological operation with the minimal bounding box) and method 6 (canny edge detection), this presented method classified as group C method, it did well in detection and consuming low time, thus it is used in the application of our thesis. The application set of our thesis consists of a prototype conveyer belt derived by a servo motor implementedespecially for our work, robot arm type Rios, and a stationary camera mounted on the top of the conveyer built. In our application five random types of objects are used, the location, size (area in pixels), and orientation for the region of interest (the objects) are detected depending on the images captured by the camera. When the object enters camera's scope, a captured image enters an image processing operation to detect the object, remark its features, and use it to make a robot arm go to the correct location to reach the object there, grip the object and move it to another location. Due to the MR results, it is obvious to notice that method 7 (Statistical Morphological Operation with MBR and Edges Detection) is gave the best results in the detection for all the used objects, according to that, method 7 is used in the implementation of the thesis application.
Exploiting temporal query behavior for improving result cache accuracy in web search engines
(Çankaya Üniversitesi, 2015) Wajji, Safaa Jamaah
In Web Search Engines responding to the user queries in a timely fashion is an important requirement. One of the integral techniques to improve the response time of a search engine is caching. By storing different types of information in a fast access memory storage, caching achieves a higher availability and better response times for the search engine. Due to anonymous and global access pattern of the queries, search engines are often considered timeless frameworks. That is, search engine sites constantly respond to queries that are submitted all around the world at an almost constant pace throughout the day. During our studies, we evaluate this phenomenon and come to the conclusion that each of the data centers, which in cooperation form the general infrastructure of a general purpose search engine in fact realizes high levels of query temporality. In this work, we aim to apply and exploit the temporal behavior of the submitted queries to improve the cache accuracy by proposing a new caching architecture. To this end, we improved the state-of-the-art result caching framework Static-Dynamic Cache (SDC) and modified it in order to incorporate query temporality. Our experiments show that the proposed caching framework improves the hit rate of a result cache up to 3%, which is roughly 25% of the possible room for improvement.

Browse

Browsing Bilgisayar Mühendisliği Bölümü Tezleri by Title