Please use this identifier to cite or link to this item: http://hdl.handle.net/11527/836
Full metadata record
DC FieldValueLanguage
dc.contributor.advisorKent, Sedeftr_TR
dc.contributor.authorİstanbullu, Mustafatr_TR
dc.date2013tr_TR
dc.date.accessioned2013-06-18tr_TR
dc.date.accessioned2015-04-21T12:00:38Z-
dc.date.available2015-04-21T12:00:38Z-
dc.date.issued2013-09-13tr_TR
dc.identifier.urihttp://hdl.handle.net/11527/836-
dc.descriptionTez (Yüksek Lisans) -- İstanbul Teknik Üniversitesi, Fen Bilimleri Enstitüsü, 2013tr_TR
dc.descriptionThesis (M.Sc.) -- İstanbul Technical University, Institute of Science and Technology, 2013en_US
dc.description.abstractBüyük bir hızla süren bilimsel ve teknolojik ilerlemeler tıp sektörüne de yansımış ve bunun sonucu olarak ileri teknoloji ürünü tıbbi cihazlar üretilmeye başlamıştır. Bu cihazlar, hastalıkların teşhisinde ve tedavi sürecinin yönetilmesi konusunda doktorlar tarafından yaygın bir şekilde kullanılmaktadır. Tıbbi görüntüleme sistemleri arasında yer alan ve vücuttan geçen X-ışınının ölçülmesi esasına dayanan bilgisayarlı tomografi cihazı; vücudun herhangi bir bölgesinin kesit görüntüsünü oluşturma kabiliyetine sahiptir. Yüksek konumsal çözünürlüğe ve doku kontrastına sahip olması sebebiyle doktorlar tarafından sıklıkla tercih edilen bir yöntem olmuştur. Bilgisayarlı tomografi cihazı, yapısı ve çalışma prensibi itibariyle kemikli dokuların incelenmesinde, yumuşak dokulara oranla daha başarılıdır. Dünya sağlık örgütünün yaptığı tanımlamaya göre kemik erimesi (aşırı kemik gözenekliliği), kemik mineral yoğunluğu azalmış ve mikromimarisi bozulmuş bir kemik hastalığıdır. Son yıllarda osteoporotik çatlama ve kırıklar toplum sağlığını tehdit eden en önemli problemlerden biri haline gelmiştir. 50 yaş ve üzerindeki kadınların menopoz sonrası dönemde bu sorunla karşılaşma oranları %30 seviyesinin üzerine çıkmaktadır. Kemik erimesi hastalığı için erken teşhis çok önemlidir. Eğer zamanında müdahale edilip tedavi aşamasına geçilmezse, hasta dokularda çatlaklara ve hatta kırıklara yol açmaktadır. Bu tez çalışmasında, bahsedilen sorunlardan esinlenerek, bilgisayarlı tomografi görüntüleri üzerinden örüntü tanıma yöntemi ile kemik erimesi olan hasta dokuların sağlıklı dokulardan ayırt edilmesi amaçlanmıştır. Çalışmanın ana yapısında dalgacık dönüşümünden faydalanılarak öznitelikler elde edilmiştir. Temel bileşen analizi yardımıyla öznitelik sayısı azaltılmış, örüntü tanımlayarak hasta olabilecek kemik dokusu ile sağlıklı kemik dokusunu sınıflandırabilmek için yapay sinir ağları ve destek vektör makineleri yöntemlerinden faydalanılmıştır. Sonuç olarak biyomedikal alanda yapılan çeşitli çalışmalarda sıkça kullanılan bu iki yöntemin başarısı karşılaştırılmış ve yapay sinir ağları ile yapılan sınıflandırmada %86 başarıyla erken dönem görüntülerine osteoporoz teşhisi konulabilmiştir. Destek vektör makineleri ile gerçekleştirilen sınıflandırmada ise doğruluk yüzdesi artırılarak %92.5 doğrulukta teşhis konulabilmektedir. Bu çalışmanın sonucunda alınmış olan başarılı sonuçlar kemik erimesi teşhisi için önemli bir katkı sağlamaktadır. 1. bölümde, çalışmanın amacı anlatılmış ve bu alanda yapılmış olan benzer çalışmalar incelenmiştir. 2. bölümde, tıpta kullanılan görüntüleme sistemlerinden kısaca bahsedilmiştir. Ayrıca bilgisayarlı tomografi sisteminin çalışma prensipleri detaylı olarak açıklanmış, kullanım alanları ve biyolojik etkilerine yer verilmiştir. 3. bölümde, kemik erimesi hastalığı üzerinde durulmuş, teşhiste kullanılan yöntemler ve konuyla ilgili bazı mühendislik problemleri detaylı olarak incelenmiştir. 4. bölüm tez çalışması kapsamında kullanılan yöntemlerin incelenmesinden oluşmuştur. Kullanılan veri setinden ve bazı ön işleme yöntemlerinden bahsedilmiş, dalgacık dönüşümü, temel bileşen analizi, yapay sinir ağları ve destek vektör makineleri detaylı olarak anlatılmıştır. 5. bölümde her iki yöntem kullanılarak elde edilmiş olan sonuçlar gösterilmiş ve son bölümde ise tüm sonuçlar yorumlanarak ilerde yapılması planlanan çalışmalardan bahsedilmiştir.tr_TR
dc.description.abstractTechnology and science develop rapidly and reflect themselves also in medical sector. Advanced technological medical devices are being produced following this. These devices are used widely by doctors in the diagnostic phase and during the management of treatment period. Computerized tomography device ranked among the medical imaging systems and based on the assessment of X-rays passing through body, has the capability of generating cross-sectional image of a part of the body. Due to its high spatial resolution and tissue contrast, it has become a frequently-preferred method among doctors. Computerized tomography device due to the nature of its structure and functioning principles is more successful in examination of bone tissues than soft tissues. According to World Health Organization, osteoporosis has been defined as a disease characterized by low bone mass and microarchitectural alterations of bone tissue, leading to enhanced bone fragility and consequent increase in fracture risk. In recent years, osteoporotic fractures or breaks have become one of the leading problems that threaten the health of society. Postmenopausal osteoporosis rate exceed 30 percent among females at the age of 50 and above. Early diagnosis is very important in osteoporosis. For beginning osteoporosis, bone loss can be slowed or prevented. A diet rich in calcium and vitamin D, or dietary supplements thereof, reduce risk of osteopenia and osteoporosis and strength-building exercise stimulates bone formation. If it is not treated or put under treatment in time, it causes cracks and even fractures in tissues with the disease. The primary goal of the diagnostic procedures is to assess the degree of bone loss for a decision on possible treatment. Whereas calcium and vitamin D supplementation are widely recommended, the type and vigorousness of a possible exercise regimen strongly depends on the degree of bone deterioration. The use of drugs also depends on the diagnosis. In advanced stages of bone deterioration it is, therefore, crucial to establish the individual fracture risk. The most common method of assessing bone strength is to monitor loss of bone mass by bone mineral densitometry. However, bone mineral densitometry is not the only factor involved in bone fragility and, therefore, in the individual risk of fracture. Indeed, considerable overlap occurs between bone mineral density values in patients with and without fractures. Other factors that influence bone strength include the bone turnover rate, bone microarchitecture, bone mass distribution, microlesion accumulation, bone crystal quality, collagen fiber quality, the degree of mineralization, and trabecular microarchitecture. The focus of recent research has been three-pronged. On the treatment side, scientists are striving to understand the cellular mechanisms that determine the balance between bone-resorbing cells (osteoclasts) and bone-forming cells (osteoblasts), with the long-term goal to influence this balance in favor of bone formation. On the diagnostic side, researchers are striving to obtain information about the bone microarchitecture, because the combined measurement of bone density and microstructural parameters promise to improve the prediction of the fracture load and therefore the individual fracture risk. Finally, basic research efforts are aimed at understanding the complex biomechanical behavior of bone. In all three cases, imaging methods play a central role. There is wide agreement that the averaging nature of the density measurement does not take into account the microarchitectural deterioration, and imaging methods that provide a prediction of the load-bearing quality of the trabecular network are actively investigated. Studies have shown that X-ray projection images, computed tomography (CT) images, and magnetic resonance images (MRI) contain texture information that relates to the trabecular density and connectivity. In this study, it is aimed to distinguish between healthy tissues and diseased tissues with osteoporosis through pattern recognition method using computerized tomography images taken from Orleans Hospital. 39 women as controls aged 67.9±9.87 standard deviation and 41 osteoporotic fractures cases aged 74±10.81 standard deviation wewre enrolled in the study. All the patiens filled out an osteoporosis risk questionnaire that included: age, personal and familial history of fracture, tobacco (yes or no), alcohol (yes or no), menopausal status, use of hormonal replacement therapy. Images were obtained on calcaneus with a direct digital computerized tomography devie. The devices for the study were cross-calibrated. Focal distance was settled at 1.15 meters. X-ray parameters were 55kV and 20mAs for all patients. Scanning the heel permitted the selection of a similar measurement site (ROI) for each subject by using anatomical landmarks. These anatomical landmarks were localized by a physician, allowing a positioning of the region of interest 1.6x1.6 cm2. The technique consist in four stages procedure. First, a high pass special frequency filter is applied to keep the essential information of the texture. The second step is the quantization of the gray level texture from 256 to 16 gray levels. In the main structure of study, image features were obtained through employment of wavelet transform. Mean and variance values of horizontal, vertical and diagonal directions are defined as features. The number of features is decreased from 6 to 2 by employing principle component analysis for each image. In the last step different models of classification are applied and compared with each other. Artificial Neural Networks and Support Vector Machines methods are used to classify healthy bone tissues and possible diseased bone tissues by pattern recognition. For artificial neural network classification, multilayer perceptron as structure and feed forward backpropagation algorithm as training method is used. As a result, the success level of these two methods which are frequently employed in various studies in biomedical field, was compared. Diagnosis of osteoporosis under the classification done with artificial neural networks was realized with the success rate of % 86 in early period images. Support vector machines also proved % 92.5 success rate in correct diagnosis of the disease. The successful results obtained in this study, provide important contributions to the diagnosis of osteoporosis. In the first chapter, the purpose of the study is explained and the similar studies conducted in this field are presented. In the second chapter, imaging systems employed in medicine are briefly described and operation principles of computerized tomography system are explained in detail and also areas of its use and biological effects are included. In the third part, osteoporosis is given special attention. Some engineering problems concerning this issue and some other methods used in diagnosis are examined in detail. In the fourth chapter of the thesis, methods used in this dissertation are examined. Dataset and some pre-process methods together with Wavelet Transform, Principal Component Analysis, Artificial Neural Networks and Support Vector Machines are explained in detail. In the fifth section, all the results obtained through use of both methods are presented. And in the last part includes interpretation of all findings and explanation of planned future studies.en_US
dc.publisherFen Bilimleri Enstitüsütr_TR
dc.publisherInstitute of Science and Technologyen_US
dc.rightsİTÜ tezleri telif hakkı ile korunmaktadır. Bunlar, bu kaynak üzerinden herhangi bir amaçla görüntülenebilir, ancak yazılı izin alınmadan herhangi bir biçimde yeniden oluşturulması veya dağıtılması yasaklanmıştır.tr_TR
dc.rightsİTÜ theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission.en_US
dc.subjectkemik erimesitr_TR
dc.subjectyapay sinir ağlarıtr_TR
dc.subjectdestek vektör makineleritr_TR
dc.subjecttemel bileşen analizitr_TR
dc.subjectdalgacık dönüşümütr_TR
dc.subjectosteoporosisen_US
dc.subjectwavelet transformen_US
dc.subjectprinciple component analysisen_US
dc.subjectartificial neural networksen_US
dc.subjectsupport vector machineen_US
dc.titleYapay Sinir Ağları Ve Destek Vektör Makineleri İle Kemik Erimesinin Teşhisitr_TR
dc.title.alternativeDiagnosis Of Osteoporosis Using Artificial Neural Networks And Support Vector Machinesen_US
dc.typeThesisen_US
dc.typeTeztr_TR
dc.contributor.authorID10003398tr_TR
dc.contributor.departmentBiyomedikal Mühendisliğitr_TR
dc.contributor.departmentBiomedical Engineeringen_US
dc.description.degreeYüksek Lisanstr_TR
dc.description.degreeM.Sc.en_US
Appears in Collections:Biyomedikal Mühendisliği Lisansüstü Programı - Yüksek Lisans

Files in This Item:
File Description SizeFormat 
13609.pdf1.92 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.