For various reasons it might be required to scan an object that partially lies outside the field of view(FOV) of a CT scanner. The parts of the object that lie outside the FOV will not contribute to the line integrals measured by the detector which will cause image artifacts that affect the final image quality. In this paper, I suggest a novel reconstruction approach that estimates the attenuation by the object outside the FOV using a priori knowledge about the outline of the object. It is shown that, knowing the object’s outline, it is possible to determine whether the attenuation along a given line is truncated. The total attenuation for a truncated projection is then estimated by interpolating the data between the consistent projections. The method therefore requires some of the projections to be consistent. This estimate, along with the knowledge of the distance traversed by the X-Ray inside the object is then used to determine the average attenuation. The method was tested on both numerical and physical phantoms. The results are satisfactory even when up to 80% of the projections are truncated. Structural Similarity Index (SSIM) was compared for the complete reconstructed images,and regions of truncations before and after the algorithm was applied. Reconstructed images from completely consistent projections served as ground truth. The results indicate that the algorithm can be used to reconstruct partially truncated CT data, which was tested on numerical and physical phantoms (of semicircular cross section). There is scope for further testing of the algorithm on irregularly shaped objects. / Technology

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.

<p dir="ltr">X-ray Computed Tomography (XCT) is a widely used non-destructive imaging technique. With advancements in hardware technology, including X-ray source and detector improvements, there is increasing demand for higher spatial resolution, faster acquisition times, and enhanced imaging quality. Moreover, customers seek better performance by addressing challenges such as metal artifacts and imaging low-attenuation objects. These evolving requirements have intensified the need to solve many ill-conditioned inversion problems inherent in X-ray imaging, driving innovations in algorithm development.</p><p dir="ltr">This thesis introduces model-based approaches to address two ill-conditioned inverse problems in X-ray imaging. In the first portion of the thesis, we address the inverse problem of estimating the effective spectrum for different X-ray systems using transmission CT measurement of homogeneous metal rods with known material and dimensions. We proposed two approaches to estimate the effective spectrum. First, we propose a dictionary-based spectral estimation (DictSE) algorithm that represents the unknown spectral response using an over-complete dictionary and finds the optimal sparse representation of the spectrum. Second, we propose a model-based spectral calibration (MBSC) algorithm that models the effective spectrum as a function of some physically meaningful parameters and estimates the spectral response by estimating a limited number of parameters with multi-voltage or multi-filtration dataset. Using simulated and measured data, we demonstrate that MBSC outperforms other methods' accuracy and robustness. </p><p dir="ltr">In the second portion of the thesis, we address the inverse problem of reconstructing images from sparse-view CT scans, a critical task in many applications to achieve faster acquisition times. We propose a direct reconstruction method called Recurrent Stacked Back Projection (RSBP), which leverages a deep recurrent neural network on the Stacked Back Projections (SBP) to improve reconstruction quality. Using simulated and experimental data, we demonstrate that our method produces more high-quality reconstructions from sparse measurements than the reconstructions from filter backprojection (FBP) and model-based iterative reconstruction (MBIR).</p>

Deux méthodes d’assistance opératoire, pour le positionnement du composant glénoïdien d’une prothèse d’épaule, sont explorées. Elles ont pour dénominateur commun une reconstruction 3D première de la glène pathologique à implanter. Une approche essentiellement clinique, avec des travaux d’application pratique, est proposée pour la technologie des Patients Specific Implants (PSI), dont l’utilisation en orthopédie est croissante. Une approche davantage technologique est ensuite proposée, de type Réalité Augmentée, jusqu’à maintenant encore inexploitée dans le champ de la chirurgie orthopédique. La faisabilité de cette approche, les conditions d’emploi des technologies inhérentes, ont été étudiées. En amont, un nouveau type d’information pour implémenter, sur le support connecté (lunettes électroniques), l’application de réalité, est proposé, avec la modélisation mathématique par régression linéaire multiple d’une glène normale. L’objectif secondaire est d’obtenir une banque de données dites de glènes génériques normales, pouvant servir de référence à la reconstitution d’une glène pathologique à traiter, après un processus de morphing. / In this thesis, two methods of operating assistance for the positioning of the glenoid component of a shoulder prosthesis, are addressed. They have in common a preliminary 3D reconstruction of the pathological glenoid to implant. A main clinical approach, with practice studies, is proposed for the Patient Specific Implants technology, which is currently used in orthopaedics. Then a main prospective and technological approach is proposed with the Augmented Reality, while it is so far untapped in the field of orthopaedic surgery. The feasibility of this last technology, as well as the tools and the manual for its use, were studied. Upstream, a new type of information to implement the augmented reality connected application support is offered, with mathematical modeling by multiple linear regression of a normal glenoid. The second goal is to build a normal generic glenoids database. It can be used as reference to the reconstruction of a pathological glenoid to treat, after a morphing process step.