Dissertations / Theses on the topic 'Tomogram images'

Дисертаційна робота присвячена розробці методів і засобів автоматизованої побудови моделі черепного імплантату та підвищенню точності планування нейрохірургічних операцій щодо реконструкції дефектів черепа. У роботі проведено огляд існуючих методів і засобів для побудови черепних імплантатів. Розглянуто основні методи обробки інтроскопічних зображень. Розроблено комплексний підхід до обробки томографічних і рентенографічних зображень голови пацієнта. Розроблено методи автоматизованої побудови об’ємних комп'ютерних моделей ЧІ за КТ- і РГ-даними з використанням 3D-моделі черепа пацієнта та усередненої моделі черепа. Розроблено метод автоматизованого визначення геометричних характеристик ЧІ, що дозволяє хірургу визначити оптимальний оперативний доступ. Розроблено медико-технічні вимоги до нейрохірургічного комплексу, призначеного для одержання вихідних даних, їхньої обробки й виготовлення речовинної копії імплантату. Проведено порівняльний аналіз отриманих результатів побудови моделей ЧІ із вже існуючими. Результати аналізу показали ефективність розроблених методів. Dissertation work considers the development of methods and facilities of the automatized construction of cranial implant model and the increase of procision of planning of neuro-surgical operative interferences for the reconstruction of cranial defects. The review of existent methods and facilities for the construction of cranial implants is provided. The basic methods of processing of introscopy images are considered. The complex approach is developed for treatment of patient’s head tomography and X-ray images. Methods of the automatized construction of three-dimensional computer models of cranial implant are developed on a base of tomography and X-ray data with the use of 3Dmodel of patient’s cranium and cranium average model. The method of the automatized determination of geometrical descriptions of cranial implant allowing a surgeon to define optimum operative access is developed. The basic medical-technical specifications are developed to neuro-surgical complex intended for the basic data obtaining, their treatment and making of implant material copy. The comparative analysis of the obtained results of models of cranial implants construction and already existing one is conducted. The results of analysis showed efficiency of the developed methods.

Thesis (Ph. D.)--Harvard University--Massachusetts Institute of Technology Division of Health Sciences and Technology, Program in Medical Engineering and Medical Physics, 1990.
Includes bibliographical references (leaves 142-150).
by Charles William Stearns.
Ph.D.

Submitted by Fabio Sobreira Campos da Costa (fabio.sobreira@ufpe.br) on 2017-08-30T12:16:52Z No. of bitstreams: 2 license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5) Dissertação - Raphael Lima Nobre de Almeida.pdf: 3877614 bytes, checksum: fb5c13466242f36f8e80a7bfdbe436ab (MD5)
Made available in DSpace on 2017-08-30T12:16:52Z (GMT). No. of bitstreams: 2 license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5) Dissertação - Raphael Lima Nobre de Almeida.pdf: 3877614 bytes, checksum: fb5c13466242f36f8e80a7bfdbe436ab (MD5) Previous issue date: 2016-03-07
O câncer de pulmão tem se mostrado o mais agressivo e mortal dentre os tipos de câncer existentes. Uma das principais formas de combatê-lo eficientemente é através do diagnóstico precoce, porém, devido ao alto número de exames de tomografia analisados, vários casos tendem a passar despercebidos. Uma das maneiras de amenizar o problema é o uso de CAD para servir como segunda opinião no momento do diagnóstico. Esse trabalho apresenta um novo sistema CAD, tendo como resultado a indicação de nódulos presentes em exames de tomografia computadorizada do tórax de pacientes. O sistema proposto se baseia em propriedades matemáticas extraídas das regiões candidatas para determinar se elas são nódulos ou não, e a classificação é realizada usando um paralelo entre classificadores baseados em aprendizagem supervisionada e não-supervisionada. Foram estudados vários métodos CAD para definição de uma estrutura de processo automática, em todas as etapas, que primeiro realiza a segmentação da imagem do pulmão a partir de exames de tomografia computadorizada, depois extrai ROI contendo áreas com potencial presença de nódulos, e depois realiza a classificação das ROIs entre áreas com Nódulos e Não-nódulos. Uma base de dados pública de imagens de exames de pacientes LIDC, internacionalmente usada na concepção e testes de sistemas CAD, foi empregada. SOM e SVM foram investigados como classificadores. Como resultado, foram obtidos valores acima de 85% de acurácia, 88% de sensibilidade, 86% F-measure e 1,59 FP/Slice para o SVM e valores acima de 81% de acurácia e 86% de sensibilidade, 82% F-measure e 2,05 FP/Slice para o SOM, onde o FP/Slice significa a média de falsos positivos gerados por imagem. O sistema proposto apresentou resultados superiores em alguns aspectos quando comparados a metodos anteriores e ligeiramente inferior quando comparado com um sistema CAD de saída semelhante, mas que realiza a segmentação manual das regiões de interesse. A partir dos resultados, foi possível ratificar a viabilidade de sistemas CAD com um novo sistema para o diagnóstico de nódulos de câncer de pulmão.
Lung cancer has demonstrated to be the most aggressive and mortal among all kinds of cancer. One of the main manners to efficiently defeat it is the early diagnostic, although due the high number of exams of computed tomography analyzed by the radiologists, they tend to let scape some true cases. A suitable way to diminish this problem is using CAD systems to be used as a second opinion during diagnostic task. Therefore, this work presents a new CAD system , showing as result the indication of the present nodules in thoracic computed tomography exams. The proposed system is based on mathematical properties extracted from Região de Interesse (Region of Interest) (ROI)s to define if they are nodules or not, the classification is accomplished using a parallel between supervised and unsupervised learning techniques.It was studied a gamut of Sistemas para Diagnóstico Assistido por Computador (Computer Aided Diagnosis) (CAD) methods to determine a struct of process, and a base of comparison between the methods. The chosen database was (Lung Image Database Consortium) (LIDC), because it is a public base, internationally used and has diagnostic to each ROI. Mapa Auto-organizável de Kohonen (Self-Organizing Map) (SOM) and Máquina de Vetor de Suporte (Support Vector Machine) (SVM) was used as classifiers, to trace a study of results among them.As result, SVM achieved more than 85% of accuracy, 86% F-measure, 88% of sensibility and 1,59 of FP/Slice. About SOM, the best parametric configuration resulted in more than 81% of accuracy, 85% of sensibility, 82% F-measure and 2,05 of FP/Slice. Thus, it was possible to deduce that the proposed system presents better score when compared with equivalents methods, and lightly worse when compared with a CAD that has ROIs segmented manually. Based on gathered results, it was possible to ratify the importance and the possibility of CAD applications, as well as presents a new system able to perform equivalent task.

Positron emission tomography (PET) is a widely used molecular imaging modality,which offers quantitative information about many biochemical processes in vivo. In particular, the dynamic PET data provide physiologically meaningful parametricimages after the estimation of the parameters of a model that best describes thekinetic behaviour of the injected radiotracer. Spatiotemporal 4D image reconstructionalgorithms estimate these physiological parameters directly from the raw sinogramdata, where the noise distribution can be more accurately modelled and thus leading tostatistically more reliable parameter estimates. In this thesis a novel direct parametricimage reconstruction algorithm is introduced, which is based on the expectationmaximisation (EM) framework and is applicable to any spatiotemporal model. Themethod is evaluated for the spectral analysis model, which is a linear temporal modeland a two-tissue compartment model, which is a nonlinear temporal model. Inaddition, the method is evaluated for a linear spatial model and in particular the modelthat is normally used to describe the blurring components in image-based resolutionmodelling. Finally, the performance of gradient-based 3D reconstruction algorithmswas also assessed as an alternative to the well-established EM-based algorithms.

Made available in DSpace on 2015-04-14T13:29:26Z (GMT). No. of bitstreams: 1 412660.pdf: 7174272 bytes, checksum: d41ce6f79c7f68c69b6ee42ff58fb76a (MD5) Previous issue date: 2008-12-18
A Tomografia Computadorizada (TC) tem ocupado lugar de destaque em trabalhos cient?ficos que demonstram sua utiliza??o cl?nica, qualidade superior de imagem e precis?o em medidas relativas aos ossos da face. No entanto, a utilidade da TC na avalia??o de anomalias craniofaciais n?o depende apenas da precis?o do m?todo, mas tamb?m da sua reprodutibilidade entre diferentes observadores. A presente pesquisa analisa a capacidade de um software biom?dico de dom?nio p?blico em delimitar o defeito ?sseo alveolar inerente ? fissura labiopalatina e mensur?-lo volumetricamente. A amostra foi constitu?da de imagens tomogr?ficas computadorizadas de vinte e um pacientes portadores de fissura labiopalatina transforame incisivo, sendo treze do g?nero masculino e oito do g?nero feminino, com m?dia de idade de dez anos e cinco meses. Os arquivos DICOM foram acessados a partir de CD-R atrav?s do software IMAGE J 1.38, instalado em computador pessoal. Ap?s a devida importa??o da seq??ncia de imagens tomogr?ficas, visualiza??o e estudo de todos os cortes, foram selecionados aqueles correspondentes ? regi?o de interesse para cada paciente. Procederam-se os delineamentos e suas respectivas mensura??es. Os resultados demonstraram um volume m?dio do defeito ?sseo alveolar inerente ? fissura de 0,77? 0,17 cm?. Na avalia??o do m?todo, a concord?ncia intra e inter-observadores foi de 95,2% e 90,5%, respectivamente. A partir da metodologia empregada e dos resultados obtidos, concluiu-se que ? poss?vel por meio do software Image J 1.38 a delimita??o nas imagens tomogr?ficas, de regi?es de interesse de estruturas anat?micas faciais e o c?lculo volum?trico da ?rea delimitada.

Submitted by Pedro Barros (pedro.silvabarros@ufpe.br) on 2018-07-19T19:32:42Z No. of bitstreams: 2 license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5) DISSERTAÇÃO Valter Augusto de Freitas Barbosa.pdf: 4679217 bytes, checksum: e249710740330a6a6d6443c6a52e880f (MD5)
Approved for entry into archive by Alice Araujo (alice.caraujo@ufpe.br) on 2018-07-20T22:18:23Z (GMT) No. of bitstreams: 2 license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5) DISSERTAÇÃO Valter Augusto de Freitas Barbosa.pdf: 4679217 bytes, checksum: e249710740330a6a6d6443c6a52e880f (MD5)
Made available in DSpace on 2018-07-20T22:18:23Z (GMT). No. of bitstreams: 2 license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5) DISSERTAÇÃO Valter Augusto de Freitas Barbosa.pdf: 4679217 bytes, checksum: e249710740330a6a6d6443c6a52e880f (MD5) Previous issue date: 2017-02-23
FACEPE
A Tomografia por Impedância Elétrica (TIE) é uma técnica de imagem não invasiva e livre de radiações ionizantes. Sua realização dar-se através de um conjunto de eletrodos dispostos na superfície do objeto ao qual se quer imagear. Pelos eletrodos é aplicado uma corrente elétrica de baixa amplitude, e em seguida, os potenciais elétricos resultantes à excitação são medidos. Em um algoritmo de reconstrução os dados de corrente e potenciais elétricos são usados para estimar a distribuição de condutividade interna do objeto à qual pode ser representada por uma imagem. O problema de reconstrução de imagens de TIE consiste na solução dos problemas direto e inverso. No problema direto é determinado os potenciais elétricos internos e de superfície da seção do objeto a partir do padrão de excitação de corrente e da distribuição de condutividade interna do domínio. Sendo tal problema, resolvido através do método dos elementos finitos. Por outro lado, a estimação da distribuição de condutividade elétrica do interior da seção do corpo a partir das medições da resposta a excitação é, matematicamente, um problema inverso, mal posto e mal condicionado. Sendo um problema complexo, de alto custo computacional, e que ainda obtém imagens de contorno suave e de baixa resolução. Uma das formas de reconstrução de TIE é através de métodos iterativos de otimização, onde o problema direto é chamado frequentemente. Este trabalho propõe a reconstrução de imagens de TIE como um método de otimização utilizando algoritmos evolucionários e bioinspirados da Inteligência Computacional tendo o Erro Médio Quadrático como função objetivo a ser minimizada. As técnicas consideradas foram: Algoritmos genéticos, Evolução Diferencial, Busca por Cardumes de Peixes e Busca por cardumes de peixes baseada em densidade, além da implementação da Busca não-cega à Busca por Cardumes de Peixes. Os algoritmos de reconstrução foram implementados em MATLAB e fazendo uso do software de código aberto EIDORS. Os experimentos foram realizados utilizando imagens padrão ouro de duas malhas de elementos finitos. De forma qualitativa as imagens obtidas foram comparadas com as imagens padrão ouro consideradas, enquanto que de forma quantitativa foi avaliado os gráficos da queda do erro em função do número de avaliações da função e pelo número de iterações dos algoritmos. As melhores imagens foram obtidas pela busca por cardume de peixes com a busca não-cega, no entanto, os menores tempos de reconstrução foram obtidos pela evolução diferencial e algoritmos genéticos.
Electrical Impedance Tomography (EIT) is a noninvasive imaging technique and free of ionizing radiation. Its implementation is given through a set of electrodes placed on object surface to be imaged. By the electrodes, a low amplitude electric current is applied, then, the resultant electric potential to the excitation is measured. In a reconstruction algorithm, the data of the current and electrical potentials are used to estimate the intern conductivity distribution of the object which it can be represented by an image. The reconstruction problem of EIT images consists in to solve the direct and inverse problems. In the direct problem is determined the intern and surface electrical potentials of the object section from the excitation pattern of electric current and the intern conductivity distribution. Such problem is resolved by the finite elements method. On the other hand, the estimation of the electrical conductivity distribution of the interior of the body section from the measures of the response to the excitation is, mathematically, an inverse problem, ill-posed and ill-conditioned. Being a complex problem, of high computational cost, and still obtains low-resolution and soft-contour images. One of the ways to reconstruct EIT images is trough iterative optimization methods, which the direct problem is used often. This work proposes the EIT image reconstruction as an optimization method using evolutionary and bioinspired algorithms from Computational Intelligence having the Root Mean Squared Error as objective function to be minimized. The techniques considered were: Genetic Algorithm, Differential Evolution, Fish School Search and Density based on Fish School Search, beyond the implementation of the Non-Blind Search to Fish School Search. The reconstruction algorithms were implemented in MATLAB using the open-source software EIDORS. Experiments were made using ground-truth images of two finite elements meshes. Qualitatively the images obtained were compared with the ground-truth images considered, whereas quantitatively were considered the graphics of the root-mean-squared error in function of the number of evaluations of objective function and in function of the number of iterations of the algorithms. The best images were obtained by the fish school search with non-blind search, however, the smallest reconstruction time were obtained by differential evolution and genetic algorithms.

Submitted by Jean Medeiros (jeanletras@uepb.edu.br) on 2017-10-23T13:03:47Z No. of bitstreams: 1 PDF - José Alberto Souza Paulino.pdf: 32306126 bytes, checksum: d8a3a4c1c537e36b743590e029308f8c (MD5)
Approved for entry into archive by Secta BC (secta.csu.bc@uepb.edu.br) on 2017-10-26T16:55:17Z (GMT) No. of bitstreams: 1 PDF - José Alberto Souza Paulino.pdf: 32306126 bytes, checksum: d8a3a4c1c537e36b743590e029308f8c (MD5)
Made available in DSpace on 2017-10-26T16:55:17Z (GMT). No. of bitstreams: 1 PDF - José Alberto Souza Paulino.pdf: 32306126 bytes, checksum: d8a3a4c1c537e36b743590e029308f8c (MD5) Previous issue date: 2017-05-08
This research proposes to evaluate and implement a solution for metal artifact reduc- tion in computed tomography, this one aiming to meet a demand from the prototyping laboratory of the Núcleo de Tecnologias Estratégicas em Saúde (Nutes) da Univer- sidade Estadual da Paraíba, where impressions of biomodels are made for surgical planning. The CT affected by metal artifacts need to be corrected prior to the printing process, this manual intervention implies excessive delay for delivery of the biomodels. The development of the proposed solution is based on the sinogram correction method which according to Mouton et al (2013) and Gjesteby (2016) is the most utilized method for reducing metal artifacts and makes uses of linear interpolation to correction the cor- rupted data. In order to validate the preference for linear interpolation in the state of the art, others interpolative techniques were implemented and evaluated; Fist through simulations and then by a form for qualitative evaluation, upon which statistical tests were applied. The results obtained confirm the use of interpolation as the best option for the reconstruction of data corrupted by metallic artifacts.
Esta pesquisa se propõe a avaliar e implementar uma solução para redução de artefatos metálicos em tomografias computadorizadas, solução esta que visa atender uma demanda do laboratório de prototipagem do Núcleo de Tecnologias Estratégicas em Saúde (Nutes) da Universidade Estadual da Paraíba, onde são realizadas impressões de biomodelos para planejamentos cirúrgicos. As tomografias afetadas por artefatos metálicos necessitam de correção antes do processo de impressão, esta intervenção realizada de forma manual implica em demora excessiva para entrega dos biomodelos. O desenvolvimento da solução proposta baseia-se no método de correção de sinograma que, de acordo com Mouton et al (2013) e Gjesteby (2016), é o método mais difundido para redução de artefatos metálicos e faz uso da técnica de interpolação linear para correção dos dados corrompidos. Objetivando validar a preferência pelo uso da interpolação linear no estado da arte, foram implementadas outras técnicas interpolativas as quais foram submetidas a avaliação; Primeiro por meio de simulações e depois via fomulário para avaliação qualitativa, na qual foram aplicados testes estatísticos. Os resultados obtidos ratificam o uso da interpolação linear como melhor opção para reconstrução de dados corrompidos pelos artefatos metálicos.

Submitted by Caroline Xavier (caroline.xavier@pucrs.br) on 2017-05-25T15:00:50Z No. of bitstreams: 1 DIS_RICARDO_GIACOMINI_DE_MARCO_PARCIAL.pdf: 425010 bytes, checksum: 70ab63ee01f40dabcc01f10861fb479f (MD5)
Made available in DSpace on 2017-05-25T15:00:50Z (GMT). No. of bitstreams: 1 DIS_RICARDO_GIACOMINI_DE_MARCO_PARCIAL.pdf: 425010 bytes, checksum: 70ab63ee01f40dabcc01f10861fb479f (MD5) Previous issue date: 2017-03-06
At the back regions of maxilla, often are unable to accommodate dental implants because of thin bone thickness. In these cases, the procedure of lifting the maxillary sinus floor is the viable option to rehabilitate this area. One of the main reasons patients give up treatment is the waiting time between the bone graft and the installation of dental implants (around 6 months). This research sought to use Human Growth Hormone (GH), together with Lyophilized Bovine Bone (Bio-Oss?), in Maxillary Sinus Floor Survey procedures to analyze whether there is a reduction in bone maturation time. In the first phase of the work, three patients were operate bilaterally, and on one side the conventional maxillary sinus survey and on the other side was add Human Growth Hormone (Somatropin). After 3 to 5 months were make Computed tomography of cases operated. The images obtained were assemble in questionnaire format, without identification of the patient, material used, time after surgery and use of GH or not. The questionnaire was forwad to 52 Oral and Maxillofacial Surgeons, by e-mail. At the end of the work, we concluded that growth hormone-enhanced grafts obtained full majority of the evaluators' agreement to install the implants at the time the Topographies were make. Unlike in the group without Growth Hormone, the great majority of the evaluators considered the moment inopportune for implant installation. These results demonstrate positive interference in the quality of the tomographic image, when the increase of the Growth Hormone was make, according to the opinion of the evaluators. Key
Regi?es posteriores de maxila, frequentemente, s?o incapazes de acomodar implantes dent?rios devido a pouca espessura de osso. Nesses casos, o procedimento de levantamento do assoalho do seio maxilar ? a op??o vi?vel para reabilitar essa ?rea. Um dos principais motivos que faz os pacientes desistirem do tratamento ? o tempo de espera entre o enxerto ?sseo e a instala??o dos implantes dent?rios (em torno de 6 meses). Esta pesquisa buscou utilizar o Horm?nio de Crescimento Humano Recombinante (rhGH), juntamente com o Osso Bovino Liofilizado (Bio-Oss?), em procedimentos de Levantamento de Assoalho de Seio Maxilar para analisarmos se h? redu??o no tempo de matura??o ?ssea. Na primeira fase do trabalho, tr?s pacientes foram operados bilateralmente, sendo que, em um lado, foi realizado o levantamento de seio maxilar convencional e, no outro lado, foi acrescentado o rhGH (Somatropina). Entre tr?s a cinco meses, foram realizadas tomografias computadorizada dos casos operados. As imagens obtidas foram montadas em formato de question?rio, sem identifica??o de paciente, material utilizado, tempo p?s-operat?rio e uso ou n?o do rhGH. O question?rio foi encaminhado para cinquenta e dois Cirurgi?es Bucomaxilofaciais e Implantodontistas por e-mail. Ao final do trabalho, conclu?mos que os enxertos com incremento do rhGH obtiveram plena maioria da concord?ncia dos avaliadores em instalar os implantes no momento em que as tomografias foram realizadas. Ao contr?rio, no grupo sem incremento do rhGH, a grande maioria dos avaliadores considerou o momento inoportuno para instala??o dos implantes. Esses resultados demonstram interfer?ncia positiva na qualidade da imagem tomogr?fica, quando realizado o incremento do Horm?nio de Crescimento, conforme a opini?o dos avaliadores.

Introduction: In Positron Emission Tomography (PET), the use of resolution modelling (RM) in iterative image reconstruction enables the modelling of aspects of detection which result in mispositioning of measured data and the subsequent blurring of reconstructed images. RM reconstruction can result in significant improvements in spatial resolution, voxel variance and count rate bias and could be a software alternative to detection hardware that is able to achieve higher resolution. Such hardware typically consists of small scintillation crystals, small bore diameters and depth of interaction discrimination, such as for the High Resolution Research Tomograph (HRRT, Siemens), which used a double crystal layer phoswich detector system. However, RM implementation comes with penalties such as slower rates of convergence, potentially higher region of interest variance and Gibbs artefacts. Methods: Assessment of the benefits and drawbacks of RM was done in the first part of this thesis together with the measurement and modelling of spatially varying resolution kernels for different scanner configurations and PET isotopes for the HRRT. It is also unclear as to whether high resolution scanning offers significant advantages over clinical PET-CT scanners for applications in the head. Through direct comparison to our HRRT, we explore whether there are significant advantages of high resolution scanning for an application in the head over clinical PET-CT. For this comparison our Biograph TruePoint TrueV (Siemens) optimised for whole body imaging was used and a novel clinical study using both scanners was set where we scanned Neurofibromatosis 2 (NF2) patients with vestibular schwannomas (VS) using [18F]fluorodeoxyglucose (FDG) and [18F]fluorothymidine (FLT). The clinical objective was to assess if uptake within VS of FLT and FDG could be measured and whether this uptake was predictive of tumour growth. Finally an assessment of the feasibility and impact of reducing the original injected activities in our clinical study was performed using bootstrapping resampling. Conclusions: RM provides greater but additive improvements in image resolution compared to DOI on the HRRT. Isotope specific image based RM could be estimated from published positron range distributions and measurements using fluorine-18. With the clinical project, uptake of FDG and FLT within the VS lesions was observed, these uptake values were correlated to each other, and high uptake was predictive of tumour growth with little differences in predictive power between FLT and FDG. Although there were benefits of the HRRT for imaging small lesions, in our clinical application there was little difference between the two scanners to discriminate lesion growth. Using the PET-CT scanner data and knowledge of lesion location, doses could be reduced to 5-10% without any significant loss of ability to discriminate lesion growth.

This paper deals with the topic of numerical arrangement of fibers in a heat exchanger. The heat exchanger is scanned in an industrial tomograph and the acquired data are represented by the field of voxels. The method used in this paper is based on tracing the fiber fragments through the use of image analysis and the subsequent numerical connection of the fragments. The result is a set of fibers that are represented by points in the field through which they are passing.

L'objectif de cette thèse étudie puis d'implémenter une méthode de reconstruction sur la tomographie aux rayons X qui permette de scanner des gros échantillons avec une bonne résolution qui permettrait de scanner l'intérieur d'un cylindre(tomographie locale) et caractériser des matériaux réels. Notre méthode peut être réalisé pour améliorer la quantitative et la qualitative d'image de fantôme. Dans cette thèse, cet méthode est utilisé sue les matériaux cellulaires. Nous avons réalisé des essais in situ en tomographie aux rayons X sur le mousse ERG et les sphères creuses. Un comparaison entre les résultats d'Élément Finis et des essais réels montre le mécanisme d'endommagement de ces matériaux.

碩士
國立成功大學
醫學工程學系
81
This study presented a CT image processing system on standard PC hardware(VGA display system) under DOS enviroment. The aim of this system is to provide a tool for medical people, who are lack of the image processing knowledge, to analyze CT image in a fast and effective manner thus can obtain more information from CT diagnosis. To achieve this fast and effective goal, this system used a special purpose and graphics user interface (GUI) approach. Because this system is designed for special purpose, most of the procedures can be made automatically. The functions provide by this system include : CT scan display (with color and gray level, definition of dynamic display density range ), automatic image segmentation, automatic inner and outer contours generation of bone in CT images, measurement of the contour's area and length between two points in a CT image, finding the locations of center of mass and centroid of bone in a CT image, calculation of moment of inertia baes on inner and outer contours of bone, and a virtual 3D view projection. The virtual 3D view can provide a reasonable 3D effect without actually creating the 3D model and is most suitable for PC in a computational standpoint.

碩士
中華大學
資訊工程學系碩士班
90
In this thesis, we present new integrated methods to segment the lung CT images and to detect the nodules more efficiently and correctly. We employed series of image processing techniques including thersholding, morphology closing, and labeling to segment the lung area and obtain the region of interests (ROIs). We extract three main features, circularity, size of area, and mean brightness from ROIs and identify the nodules with pure fuzzy inference rules and neural fuzzy model. Twenty-nine clinical cases involving 583 images (512X512X8 bits) were tested in our study. The detection rate of the proposed methods are 89.3%, and the false positives are approximately 0.3 and 0.21 per image respectively. This result demonstrates that our method improves the detection rate and reduces false positive compaired to other approaches. The studies have shown high potential implementation of this system in clinical practice.

碩士
國立中央大學
機械工程研究所
99
Ultrasound　imaging　is　a　convenient　tool　for　diagnosis　of　abdominal　organ　diseases　and　can　be　applied　to　biopsy.　Ultrasound　imaging　is　real-time　and　has　no　harm　to　human　body.　However　it　has　some　drawbacks　such　as　images　are　blurred　and　lack　of　spatial　information.　　Physicians　need　well　training　and　clinic　experience　to　avoid　misdiagnosis.　This　study　proposes　an　image　registration　method　between　ultrasound　images　and　CT　images,　which　can　be　used　to　develop　an　ultrasound　scan　training　program　for　medical　staffs. 　　First,　ultrasound　images　and　corresponding　spatial　information　are　captured　by　computer　program,　and　then　the　images　are　calibrated　by　using　an　optic　tracker　and　a　calibration　phantom　with　N-type　wires.　The　registration　between　CT　images　and　US　images　is　done　by　finding　the　transformation　matrix　using　spherical　markers　combined　with　optimization　method　–　the　Iterative　Closest　Point　algorithm.　Then,　the　CT　and　US　images　can　be　fused　to　build　a　training　program　using　MCV　software　architecture.　The　registration　errors　can　also　be　assessed　in　the　training　program. 　　Registration　error　assessment　experiments　were　performed　with　a　box　phantom,　an　anthropomorphic　phantom,　and　patients.　The　errors　using　the　box　phantom　and　anthropomorphic　phantom　were　about　1-2　mm.　But　the　results　of　patient　experiments　had　an　error　of　1-2　cm.　The　possible　main　factor　is　liver　movement　due　to　respiration　which　should　be　compensated.

A method for calculating Fourier coefficients of functions of three variables using tomograms is proposed and investigated. Earlier Lytvyn O.M. proposed and introduced the method of calculating Fourier coefficients of the function of two variables using integrals of this function along a given system of direct. It is believed that these integrals - projections or projection data arrive from a computer tomograph.

Accurate and automatic segmentation techniques that do not require any explicit prior model have been of high interest in the medical community. We propose a fully-automatic method for segmenting the femur from 3D Computed Tomography scans, based on the graph-cut segmentation framework and the bone boundary enhancement filter analyzing second-order local structures. The presented algorithm is evaluated in large-scale experiments, conducted on 197 CT volumes, and compared to other three automatic bone segmentation methods. Out of the four tested approaches, the proposed algorithm achieved most accurate results and segmented the femur correctly in 81% of the cases.

碩士
國立中央大學
機械工程研究所
100
Ultrasound scan is the most common way to detect the disease of liver diseases. Ultrasound imaging is real-time and non-invasive, but its image quality is poor and lack of spatial information. Therefore, subjective interpretation of ultrasound images may cause misdiagnosis. On the other hand, computed tomography images have high quality. If CT and US images are scanned at similar respiration and posture status, the fusion of both images will be able to improve diagnosis quality and biopsy accuracy. Prior to ultrasound scan, the patient''s posture and respiration are monitored and controlled to have similar status as CT scan. Then the grabbed ultrasound images of ribs are processed to find the boundary of the ribs. Using the iterative closet point algorithm, the data of boundary points are then applied to complete surface registration with reconstructed 3D model of rib’s CT images and to obtain the transformation matrix between the two image frames. Experiments of image registration error assessment are performed with an anthropomorphic phantom and a human patient. The edge information of captured characteristics ultrasound images are transformed to CT image frame and compared with the same characteristics on the CT images to calculate their position errors. The results show that position error of anthropomorphic phantom experiment at the xy plane is 2.32mm and the z axis is 4.48mm while that of human patient is approximate 4mm. Since there is clear angular error, press of ultrasound probe on the patient during ultrasound scan may be the factor caused angular error.

Dissertação de Mestrado em Engenharia Biomédica apresentada à Faculdade de Ciências e Tecnologia da Universidade de Coimbra.
Cancer is the leading cause of death in the World and radiotherapy one of the main procedures for treatment. Modern radiotherapy techniques provides increasingly higher conformality to the tumor. However, for successful treatment and to ensure maximum dose exposure to the tumor and avoid the irradiation of organs at risk or healthy tissues it is important detect patient misalignments and anatomical morphological alterations before the treatment, so to carry out the necessary corrections. In this way, a new imaging system to be applied to external X-ray beam radiotherapy was proposed: OrthoCT (orthogonal computed tomography). This technique focusses on acquiring images of the volume to be irradiated immediatly before treatment, when the patient is already positioned and xed for treatment (\on-board imaging") with very low to eventually null dose administered to organs at risk or healthy tissues even when position close to the tumor. OrthoCT relies on orthogonal ray imaging, namely, the detection of scattered photons in the patient that are emitted perpendicularly to the incident beam direction. OrthoCT shows good correlation with the morphological structure of the patient. For that, it is necessary that the beam has an inferior size ( 5 mm x 5 mm) when compared to the typical ones used in patient irradiation. Moreover the scanned pro le requires: (1) high homogeneity so that variations can be associated only to dose/morphological alterations; and (2) high velocity (which favors multi-leaf collimator-based scans in respect to jaw-based scans). The purpose of this dissertation is to de ne the best scan to be applied in OrthoCT, considering the conditions aforementioned. For that some small-beam scans were evaluated experimentally and the obtained pro les were optimize. The more homogeneous pro le achieved experimentally reveals an intensity variability inferior to 1%, thus substantiating the capability of OrthoCT to detect such morphological variations or larger.
O cancro e uma das principais causas de morte no Mundo e a radioterapia um dos m etodos de tratamento mais utilizados. As t ecnicas modernas de radioterapia permitem aumentar a conformacionalidade ao tumor. Todavia, para o sucesso da terapia e de modo a garantir uma dose m axima no tumor e evitar a irradia c~ao de org~aos de risco ou tecidos s~aos adjacentes, e importante detetar os desposicionamentos do doente e as altera c~oes anat omico-morfol ogicas antes do tratamento ter in cio, para que se possa proceder as devidas corre c~oes. Nesse sentido, um novo sistema de imagem para radioterapia de feixe externo com raios X foi proposto: OrthoCT (orthogonal computed tomography). Este sistema pretende adquirir imagens do volume a ser irradiado imediatamente antes do tratamento ter in cio, quando o doente j a se encontra posicionado e imobilizado para receber o tratamento (\on-board imaging"). Esta t ecnica e poss vel administrando uma dose muito baixa, eventualmente nula, nos org~aos de risco e tecidos saud aveis, mesmo quando posicionados pr oximos do tumor. A OrthoCT tem por base a imagiologia por raios ortogonais, isto e, a dete c~ao da radia c~ao dispersa no doente que escapa perpendicularmente a dire c~ao do feixe de irradia c~ao, pois esta apresenta uma boa correla c~ao com a estrutura morfol ogica do doente. Para isto, o feixe de irradia c~ao deve ter um tamanho inferior ( 5mm 5mm) aos tipicamente usados nos tratamentos atuais. Al em disso, o per l do varrimento deve ser: (1) o mais homog eneo poss vel, para que as varia c~oes possam ser associadas a altera c~oes de dose/morfol ogicas; e (2) ter elevada velocidade (o que favorece os varrimentos com o colimador multi-folhas, em vez das mand bulas). O objetivo deste projeto e determinar o melhor varrimento a ser usado experimentalmente em OrthoCT tendo em conta as condi c~oes referidas anteriormente. Para tal, utilizaram-se feixes nos e os resultados dos varrimentos obtidos experimentalmente foram optimizados. O varrimento mais homog eneo obtido experimentalmente revela uma variabilidade da intensidade do feixe inferior a 1 %, o que suporta altera c~oes de variabilidade morfol ogica detet aveis atrav es da t ecnica OrthoCT desta ordem de grandeza ou superior.