3D Reconstruction And Visualization Of Medical Images Based On Rational Gaussian Surfaces

The development and application of medical imaging equipments makes anatomic sturcutre of human can be acquired non-invasively, which greatly promots the development of iatrology. Usually what they can acquire are slice images of human, from which relative space position relations between tissues are not easy to get for users. Yet, tissues and pathological parts of the human body can be reconstructed and showed in three-dimensional from by visualization technology, which can provide more useful information for clinicians.Therefore, research of visualization technology applied to medical images is of great significance.This thesis introduces the survey and development trend of 3D medical visualization, the key technology and algorithm in this field are discussed and compared. Then the research scheme has been confirmed. In this paper, a 3D reconstruction and visualization of medical Images based on rational gaussian surface method is introduced that enables the user to effectively complete surface approximation and edit the surface shape by changed the position of control points.First, we approximation a digital shape with a triangular mesh. This major includes: subdivision of a single edge algorithm, subdivision of a single triangle algorithm, approximation of a digital shape by a triangular mesh algorithm. Then, approximation the triangular mesh model based on rational gaussian surface. This major includes: the principle and generation methods of rational gaussian surfaces, using the results of meshing to strike the node parameter, give the smoothing parameter selection method and completed the algorithm of approximating digital shape. Both triangulation and surface approximation proceed from coarse to fine. At last, the method implements the 3-D visualization in Visual C++.NET platforms with OpenGL, and gives Three-dimensional reconstruction system.The user can edit the surface shape by changed the parameters of the position of control points and complete editing 3D reconstruction results functions. In this thiese, because the data sources are the automatic segmentation of MRI images, it can be corrected segmentation errors .With this method, the user typically needs only a few minutes to grid subdivision,three-dimensional surface reconstruction and edit the surface shape.