الفهرس | Only 14 pages are availabe for public view |
Abstract The human knee joint has been the subject of investigation from many points of view as it is an important component of the human motion system. Computer Tomography(CT) images are usually evaluated by human evaluator in the medical praxis, however, this can be extremely difficult due to the similar gray scale representation of the synovial fluid between the opposite cartilage surfaces of femur, tibia and the cartilage covering surfaces. Efficient and sensitive computer methods for contour detection and segmentation can improve the quality of evaluation. Accurate computation of distances and directions within the joint provides solid basis for orthopedic handling and surgery. Many important properties of the knee joint can only be evaluated in three dimensions. The bone and cartilage surfaces have complicated shapes in 3D and the motion takes place in 3D. Consequently, 3D computer programs built models using information extracted from the 2D images. The purpose of this thesis is to offer a convenient geometrical modeling in a visual system to investigate the human knee joint by computer tools. The proposed model consists of the following system: input data, storing & handling of CT images of the knee, methods & programs for image analysis, 3D surface reconstruction, graphical & geometrical tools for the detailed analysis of the shape and the contact path motion of the knee. Most important elements of the system are reported in this thesis. New algorithms and computer programs are represented herein to predict the point clouds by slice adaptive threshold from CT images. These algorithms handle imperfections coming from contour detection and measurements. Then a continuous surface fit starts with topological ordering of points, which is done by creating a mesh grid from point clouds. Mesh grid provides neighborhood relations, which is important for fitting surfaces and for geometric calculations. Also a 3D model for the reconstruction of human knee joint represented by Bezier surface and polynomial regression is achieved and discussed. Finally a computer aided geometric model is used to develop and predict the cartilage and menisci of the knee. An accurate description of the total knee replacement is achieved. |