Thesis (MSc (Mathematical Sciences))--Stellenbosch University, 2008. / A novel implementation exploiting modern hardware is explored and found to be a significant improvement over current methods used. A 50 times performance increase in the computation time of DRRs is achieved over the conventional ray casting approach and image registration is performed in under a minute.

Lung cancer is a disease affecting millions of people every year and poses a serious threat to global public health. Accurate lung cancer staging is crucial to choose an appropriate treatment protocol and to determine prognosis, this requires the acquisition of contrast-enhanced diagnostic CT (d-CT) that is usually followed by a PET/CT scan. Information from both d-CT and PET scan is used by the clinician in the staging process; however, these images are not intrinsically aligned because they are acquired on different days and on different scanners. Establishing anatomical correspondence, i.e., aligning the d-CT and the PET images is an inherently difficult task due to the absence of a direct relationship between the intensities of the images. The CT acquired during the PET/CT scan is used for attenuation correction (AC-CT) and is implicitly aligned with the PET image as they are acquired at the same time using a hybrid scanner. Patients are required to maintain shallow breathing for both scans. In contrast to that, the d-CT image is acquired after the injection of a contrast agent, and patients are required to maximally inhale, for better view of the lungs. Differences in the AC-CT and d-CT image volumes are thus due to differences in breathhold positions and image contrast. Nonetheless, both images are from the same modality. In this thesis, we present a new approach that aligns the d-CT with the PET image through an indirect registration process that uses the AC-CT. The deformation field obtained after the registration of the AC-CT to d-CT is used to align the PET image to the d-CT. Conventional image registration techniques deform the entire image using homogeneous regularization without taking into consideration the physical properties of the various anatomical structures. This homogeneous regularization may lead to physiologically and physically implausible deformations. To register the d-CT and AC-CT images, we developed a 3D registration framework based on a fluid transformation model including three physically motivated properties: (i) sliding motion of the lungs against the pleura; (ii) preservation of rigid structures; and (iii) preservation of topology. The sliding motion is modeled using a direction dependent regularization that decouples the tangential and the normal components of the external force term. The rigid shape of the bones is preserved using a spatially varying filter for the deformations. Finally, the topology is maintained using the concept of log-unbiased deformations. To solve the multi-modal registration problem due to the contrast injected for the d-CT, but lack thereof in the AC-CT, we use local cross correlation (LCC) as the similarity measure. To illustrate and validate the proposed registration framework, different intra-patient CT datasets are used, including the NCAT phantom, EMPIRE10 and POPI datasets. Results show that our proposed registration framework provides improved alignment and physically motivated deformations when compared to the classic elastic and fluid registration techniques. The final goal of our work was to demonstrate the clinical utility of our new approach that aligns d-CT and PET/AC-CT images for fusion. We apply our method to ten real patients. Our results show that the PET images have much improved alignment with the d-CT images using our proposed registration technique. Our method was successful in providing a good overlap of the lungs, improved alignment of the tumours and a lower target registration error for landmarks in comparison to the classic fluid registration. The main contribution of this thesis is the development of a comprehensive registration framework that integrates important physical properties into a state-of-the-art transformation model with application to lung imaging in cancer.

Modern medicine has become reliant on medical imaging. Multiple modalities, e. g. magnetic resonance imaging (MRI), computed tomography (CT), etc. , are used to provide as much information about the patient as possible. The problem of geometrically aligning the resulting images is called image registration. Mutual information, an information theoretic similarity measure, allows for automated intermodal image registration algorithms. <br /><br /> In applications such as cancer therapy, diagnosticians are more concerned with the alignment of images over a region of interest such as a cancerous lesion, than over an entire image set. Attempts to register only the regions of interest, defined manually by diagnosticians, fail due to inaccurate mutual information estimation over the region of overlap of these small regions. <br /><br /> This thesis examines the region of union as an alternative to the region of overlap. We demonstrate that the region of union improves the accuracy and reliability of mutual information estimation over small regions. <br /><br /> We also present two new mutual information based similarity measures which allow for localized image registration by combining local and global image information. The new similarity measures are based on convex combinations of the information contained in the regions of interest and the information contained in the global images. <br /><br /> Preliminary results indicate that the proposed similarity measures are capable of localizing image registration. Experiments using medical images from computer tomography and positron emission tomography demonstrate the initial success of these measures. <br /><br /> Finally, in other applications, auto-detection of regions of interest may prove useful and would allow for fully automated localized image registration. We examine methods to automatically detect potential regions of interest based on local activity level and present some encouraging results.

Nonrigid image registration has received much attention in the medical imaging and computer vision research communities, because it enables a wide variety of applications. Feature tracking, segmentation, classification, temporal image differencing, tumour growth estimation, and pharmacokinetic modeling are examples of the many tasks that are enhanced by the use of aligned imagery. Over the years, the medical imaging and computer vision communties have developed and refined image registration techniques in parallel, often based on similar assumptions or underlying paradigms. This thesis focuses on variational registration, which comprises a subset of nonrigid image registration. It is divided into chapters that are based on fundamental aspects of the variational registration problem: image dissimilarity measures, changing overlap regions, regularizers, and computational solution strategies. Key contributions include the development of local versions of standard dissimilarity measures, the handling of changing overlap regions in a manner that is insensitive to the amount of non-interesting background information, the combination of two standard taxonomies of regularizers, and the generalization of solution techniques based on Fourier methods and the Demons algorithm for use with many regularizers. To illustrate and validate the various contributions, two sets of example imagery are used: 3D CT, MR, and PET images of the brain as well as 3D CT images of lung cancer patients.

Följande kandidatuppsats undersöker hur en elektronisk hälsoapplikation kan utformasför att öka medvetenheten om den personliga dagliga hälsan, fokuserat på huden och atthålla den frisk. Det sker genom att en prototyp har utformats där användaren genombilder samt text får instruktioner för att utföra en självkontroll. Prototypen innehållerockså en bildregistrering där användaren kan ladda upp bilder på en leverfläck ochgenom registreringen kan om leverfläcken utvecklas över tid. Syftet med prototypen äratt undersöka om en elektronisk hälsoapplikation utformad med fokus på återkopplingkan göra det möjligt för en person som inte är utbildad inom ämnet att undersöka sinhud samt se och följa förändring i en leverfläck genom bildanalys. Detta undersöktesgenom två tester. Ett frågeformulär där svarspersonen fick försöka avgöra om enleverfläck var frisk eller inte. Ett användartest där testpersonen fick testa bildanalysen,det som testades var om personen kunde ladda upp en bild samt se skillnad i bildernagenom bildregistreringen.Resultatet visar att majoriteten av enkätsvaren var korrekta svar, närmare bestämt 76,9procent. Personerna kunde identifiera om leverfläcken var frisk eller sjuk med hjälp avinformationen som tilldelades. Resultatet av användartesten visade att personerna kundeladda upp en bild på webbplatsen och sedan förstå bildanalysen. / This bachelor thesis examins how a electronic health application can be made toincrease awareness about personal daily health, with focus on individual skin andkeeping it healthy. This was made by creating a prototype where the user get usefulinformation on how to perform a self examination through text and images. Theprototype also contains functionality that allows the user to upload images of a moleand by an image registration follow possible changes over long periods of time. Thepurpose of the prototype is to investigate if an electronic health application with focuson feedback can make it possible for a person who is not educated in this area to examinthe skin and also discover possible changes over time with the image registration. Twotests were made to examin if this was possible. The first test was an questionnaire wherethe respondet tried to determine if a mole was healthy or not. The second test was a usertest where the test person got to try out the functionality with image registration in theprototype. The person got to upload an image and see the differences through the imageregistration.The results showed that the majority of the collected repsonses from the quetsionnairewas correct, with a percentage of 76,9 correct answers. The respondent could identify ifthe mole was healthy or not through the information that was presented. The results ofthe user test showed that all test persons could perform and understand the functionalityof uploading an image and then analyze the results throgh image registration.

Vitamin D is very important for endochondral ossification and it is commonly insufficient during pregnancy (Javaid et al., 2006). Insufficiency of vitamin D during pregnancy predicts bone mass and hence predicts adult osteoporosis (Javaid et al., 2006). The relationship between maternal vitamin D and manually measured fetal biometry has been studied (Mahon et al., 2009). However, manual fetal biometry especially volumetric measurements are subjective, time-consuming and possibly irreproducible. Computerised measurements can overcome or at least reduce such problems. This thesis concerns the development and evaluation of novel methods to do this. This thesis makes three contributions. Firstly, we have developed a novel technique based on the Random Forests (RF) classifier to segment and measure several fetal femoral characteristics from 3D ultrasound volumes automatically. We propose a feature selection step in the training stage to eliminate irrelevant features and utilise the "good" ones. We also develop a weighted voting mechanism to weight tree probabilistic decisions in the RF classifier. We show that the new RF classifier is more accurate than the classic method (Yaqub et al., 2010b, Yaqub et al., 2011b). We achieved 83% segmentation precision using the proposed technique compared to manually segmented volumes. The proposed segmentation technique was also validated on segmenting adult brain structures in MR images and it showed excellent accuracy. The second contribution is a wavelet-based image fusion technique to enhance the quality of the fetal femur and to compensate for missing information in one volume due to signal attenuation and acoustic shadowing. We show that using image fusion to increase the image quality of ultrasound images of bony structures leads to a more accurate and reproducible assessment and measurement qualitatively and quantitatively (Yaqub et al., 2010a, Yaqub et al., 2011a). The third contribution concerns the analysis of data from a cohort study of 450 fetal femoral ultrasound volumes (18-21 week gestation). The femur length, cross-sectional areas, volume, splaying indices and angles were automatically measured using the RF method. The relationship between these measurements and the fetal gestational age and maternal vitamin D was investigated. Segmentation of a fetal femur is fast (2.3s/volume), thanks to the parallel implementation. The femur volume, length, splaying index were found to significantly correlate with fetal gestational age. Furthermore, significant correlations between the automatic measurements and 10 nmol increment in maternal 25OHD during second trimester were found.

In radiation therapy, it is imperative to deliver high doses of radiation to the tumor while reducing radiation to the healthy tissue. Respiratory motion is the most significant source of errors during treatment. Therefore, it is essential to accurately model respiratory motion for precise and effective radiation delivery. Many approaches exist to account for respiratory motion, such as controlled breath hold and respiratory gating, and they have been relatively successful. They still present many drawbacks. Thus, research has been expanded to tumor tracking. The overall goal of 4D-CT is to predict tumor motion in real time, and this work attempts to move in that direction. The following work addresses both the temporal and the spatial aspects of four-dimensional CT reconstruction. The aims of the paper are to (1) estimate the temporal parameters of 4D models for anatomy deformation using a novel neural network approach and (2) to use intelligently chosen non-uniform, non-separable splines to improve the spatial resolution of the deformation models in image registration.

V první části se práce se zabývá rešerší metod a principů potřebných při registraci obrazu. Dále pak popisuje zobrazovací systémy očního pozadí jako jsou OCT, fundus kamera a SLO. Druhá část práce je zaměřena na praktickou realizaci programů pro registraci snímků z OCT, SLO a fundus kamery v programovém prostředí Matlab.

This document deals with the pre-processing and the segmentation of human vertebrae from computed tomography 3D image data. It describes image registration and segmentation methods concretely and commonly used in image data processing during the automated spine canal localization process and the localization of each vertebrae centroid. The 3D deformable model development is described in the work together with the introduction to the previous state of art and the progress of vertebrae segmentation using 3D deformable models registration. This work presents the evaluation of the accuracy and the reliability from the segmentation and localization results.

Durante um procedimento cirúrgico cerebral existe o deslocamento das estruturas que é um problema tipicamente não-rígido e não-linear. A ultrassonografia intra-operatória é utilizada como guia cirúrgico e pode ser utilizada para correção das imagens pré- operatórias através do corregistro rígido entre estas e um sistema de rastreio. Isto torna possível a visualização do deslocamento das estruturas devida a remoção de parte delas durante o ato cirúrgico. O objetivo deste trabalho é um estudo do corregistro livre não-rígido a partir de um modelo in vitro experimental que simule uma situação cirúrgica de retirada de uma inclusão líquida, de forma controlada, para medir os deslocamentos das estruturas próximas, utilizando imagens de ultrassom. Alguns fantomas que simulam o tecido humano nas imagens de ultrassom, feitos de gelatina e parafina, foram escolhidos como modelo. Para realizar o corregistro foi escolhida a transformação geométrica por splines simples (B-Splines), o otimizador Limited- memory BroydenFletcherGoldfarbShanno (LBFGS) e a métrica de similaridade soma do quadrado das diferenças (SQD) e, utilizada a biblioteca Insight Segmentation and Registration Toolkit (ITK), assim como o estudo dos parâmetros adequados para a nossa tarefa. Foi demonstrado para as condições envolvidas que para as imagens em modo B as deformações até 5% e mapas de RF até 9%, sem nenhuma otimização dos parâmetros do corregistro, é factível sem uso excessivo de tempo computacional. Foi analisada a influência da grade em relação a dois tipos diferentes de deformação, ambas com valor de 2%. O tamanho da grade, levando em consideração o erro e o tempo, foram a 5x11 para as imagens em Modo B e 11x17 para os mapas de RF, independentemente do tipo de deformação. Os parâmetros do otimizador (Default Step Length, Gradient Convergence Tolerance e Line Search Accuraccy) também foram avaliados e os valores obtidos foram 1,6; 0,03 e 0,8 para as imagens modo B e 1,2; 0,05 e 1,0 para os mapas de RF. No entanto ao comparamos, utilizando os parâmetros propostos obtidos, os campos de deslocamentos esperados com os gerados pelo modo B e pelos mapas RF, foi demonstrado que os mapas de RF fornecem valores abaixo do esperado e que as imagens em modo B retratam mais fielmente os deslocamentos e isto se deve a escolha do conjunto de valores testados para o otimizador. Foram aplicados estes parâmetros em dois fantomas de parafina- gel e em dois de gelatina. Nos três primeiros fantomas foi retirada um inclusão líquida em várias etapas. Os deslocamentos das estruturas vizinhas foram avaliados durante as etapas de remoção para demonstrar os campos de sução e de torção. No último fantoma, que simula morfologicamente um cérebro humano, foram retiradas, em várias etapas, regiões sólidas, simulando a retirada de tecido e foram calculados os deslocamentos e demonstrados os campos provenientes deste tipo de intervenção. Os trabalhos futuros se concentrarão em utilizar os volumes para medir os movimentos das estruturas e em novos parâmetros do otimizador para os mapas de RF. / During a brain surgery there is the displacement of the structures that is a typical non- rigid and non-linear problem. Intraoperative ultrasound is used as a surgical guide and can be used for spatial correction of preoperative images through the rigid registration between these and a track system. This makes it possible to visualize the displacement of structures due to removal of some piece of them during surgery. This work is a study of the non-rigid free-from registration using an experimental in vitro model to simulate a surgical situation withdrawal of a fluid inclusion in a controlled manner, to measure the displacement of nearby structures, using ultrasound images. Some phantoms that simulate the human tissue in the ultrasound images made of gelatin and paraffin were chosen as a model. To perform the registration it was used the framework Insight Segmentation and Registration Toolkit (ITK) and were chosen a geometric transformation of simple splines (B-splines), the Limited-memory Broyden-Fletcher- Goldfarb-Shanno (LBFGS) optimizer and the similarity metric sum of the squared differences (SQD). The search for the suitable parameters for our task are done and it has been shown that for the conditions involved for B-mode images deformations up to 5% and RF maps up to 9% without any optimization of the parameters of registration, is feasible without excessive use of computational time. The influence of the grid was examined for two different types of deformation, both for 2%. The size of the grid, taking into account the error and time were the 5x11 for the images in B mode and 11x17 maps for RF, regardless of the type of deformation. The parameters of the optimizer (Default Step Length, Gradient Convergence Tolerance and Line Search Accuraccy) were also evaluated and the values obtained were 1.6, 0.03 and 0.8 for the B-mode images and 1.2, 0.05 and 1.0 for RF maps. However when comparing the expected displacement fields with the generated by B-mode images and the RF maps, using the obtained parameters, it have been shown that RF maps provide values are lower than expected and that the B-mode images portray more faithfully displacements. This is due to the choice set of values tested for the optimizer. Finally, image registration parameters for B-mode were applied in two paraffin-gel and two gelatin phantoms. In the first three phantoms the fluid inclusion was removed in several stages and the displacements of neighboring structures were evaluated during the removal steps to demonstrate the fields of suction and torsion. The last phantom, which morphologically mimics a human brain, a solid region was removed, also in several stages, simulating a surgery. The displacements were calculated and demonstrated the fields from this type of intervention. Future work will focus on using the volumes to measure the movements of the structures and new parameters test of the optimizer to RF maps.