Registration for the in-vivo studies of osteoarthritis based on magnetic resonance imaging

Osteoarthritis (OA) affects cartilage and bones of weight bearing joints. To understand OA, assessments of joint properties and health status are needed. For these analyses, magnetic resonance imaging (MRI) can provide accurate in-vivo 3D surfaces of joint structures. Alignment of these surfaces through a registration process can allow direct comparisons between datasets. Registrations are commonly used in Geomatics engineering where temporal geographic data are compared for change detection. Therefore, the main objective of this thesis is to translate a Geomatics algorithm to register MR joint surfaces for quantitative studies of joint conditions. Experiments with Geomatics and MRI data confirmed that the algorithm could successfully register and detect discrepancies between the surfaces. Validation and repeatability studies showed that the algorithm achieved an accuracy of an image pixel size and the digitization and registration processes were highly repeatable. Applications' results also confirmed the feasibility of the algorithm for in-vivo studies of OA.