| The major goal of this dissertation is to develop a set of techniques for automatically extracting vertebral segments from three-dimensional computerized tomography (CT) and magnetic resonance (MR) images of the human cervical spine. An important motivation for this work is to provide accurate information for registration and for fusion of CT and MR images into a composite three-dimensional image. One of the major hurdles in performing image fusion is the difficulty of extracting and matching corresponding anatomical regions in an accurate, robust, and timely manner. The complementary properties of soft and bony tissues revealed in CT and MR imaging modalities makes it challenging to extract corresponding regions that can be correlated in an accurate and robust manner. Ambiguities in the images due to noise, distortion, limited resolution, and patient-specific structural variations also create additional challenges. Whereas fusion of CT and MR images of the cranium have already been performed, no one has yet developed an automated technique for fusing multimodality images of the spine. Unlike the head, which is relatively rigid, the spine is a complex, articulating object and is subject to structural deformation throughout the multimodal scanning process.; The hypothesis of my research is that by using finite-element techniques and model-based segmentation algorithms, it is possible to extract vertebral segments and determine anatomical landmarks robustly from three-dimensional CT image sets of the human cervical spine, to an accuracy of 95 percent or better. The accuracy is defined as the average sensitivity and specificity of the extracted vertebrae when compared against a gold-standard reference. The original contribution of this dissertation is to demonstrate the feasibility of a unified approach to visualization, modeling, and perception using finite-element techniques and optimization methods.; Although the techniques proposed here have been designed to work with both CT and MR images, I have only performed a rigorous test of the validity of the hypothesis on CT images. Whereas the performance of these techniques on MR images is of significant interest, I have chosen not to perform such an evaluation. A major contribution of this research is the construction of a generic finite element geometric model of the cervical spine. |
