Research On The Methods Of Reconstruction Techniques In Medical Image Visualization

3-D Reconstruction from Medical Images, a new subject, developed from 1980s, of which with the Computer Graphics and Image Processing Technology, calculation process and calculation results are converted into graph and images , displayed on the screen and interactive processing treated. In this way, processing speed and quality are improved greatly. Thus 3-D Reconstruction from Medical Images is of profound practical application background, which is widely used in the fields of medical diagnosis, surgical planning, simulation and radiotherapy programming. Therefore the research of the visualization of medical image with wonderful foreground has great academic significance and application value.The research of 3-D Reconstruction from Medical Image includes pretreatment of input image, image segmentation, model construction and rendering etc.Trying different reconstruction methods is conducive to solve bottleneck problem in medical science.In this thesis, rendering methods and development platform are discussed according to the implementation methods of 3-D Reconstruction from Medical Image. The experiment of liver edge extraction is done according to the face reconstruction method from paper[42]. Then using MATLAB and image processing method, reading and writing of the liver CT image format conversion and Image Enhancement are treated, which the segmentations of contour curve from 2D image are obtained by the structural characteristics of image. Particularly, the 3D images of liver are reconstructed and the contour curves is obtained while 20-order Fourier series are used for the image fitting.For VTK+MFC, the MarchingCubes algorithm of surface rendering 3D reconstruction based on VTK (Visualization ToolKit) is the programmed. Midpoint algorithm, which avoids the traditional linear difference, is introduced to calculate coordinates of the equivalent points and normal vector, meanwhile, the efficiency of the reconstruction is improved, iso-surfaces are smoothed, and good reconstruction effect is received. For MITK (Medical Imaging Toolkit) + MFC, Ray casting algorithm of 3D reconstruction is programmed, by which 3D images and vessels of the liver are reconstructed from DICOM format of CT images. At the same time, the factors which affect the result of the volume rendering are analyzed and illustrated.Additionally, the optical model of the surface rendering and volume rendering, geometric transformation (rotation, scaling), creating of the stereomodel and the improvement of the Marching Cubes algorithm are discussed, and realized in the software system.. Specially, data types in the data field, the method of the iso-surface smoothing are studied. Finally, the indexes and results of the reconstruction methods mentioned above are summarized and compared, from which we know that the method of the MITK volume rendering can reconstruct the liver vessels, but the performances of rapidity and interaction are poor, on the contrary, using the method of VTK surface rendering can reconstructed rapidly and interacted but can't reconstruct the structure of internal organs, and both rapidity and interaction are not owned by MATLAB surface rendering.