Image Visualization Algorithm Based On Surface Rendering Study

With the great development of both image acquiring and computing techniques, radiology and the medical image processing technique derived from it have become increasingly important in medical applications. When the scientific visualizing technique, a branch of the computer graphics, has gradually developed with many successful methods and algorithms, the three-dimensional (3D) reconstruction of medical images has become a helpful means for clinic diagnosis and treatments. On the other hand, the increasing size of the applied data sets also speeds up the development of scientific visualization.This dissertation focuses on the 3D visualization of medical images based on surface rendering. Grounds on analysis of the surface rendering methods, refinements to both the modeling and the rendering processes are presented. The main points of this dissertation are as follows:1. The principle and implementation of the Marching Cubes, a well-known surface extraction algorithm, are demonstrated, and the possible ambiguity of the MC algorithm with its solutions is proposed.2. A refinement to MC is applied, which avoids the access of uninterested cubes, also known as the'empty cubes'. Results of experiments on different data sets show the refinement accelerates the marching.3. The simplification of the surface model using the Quadric Error Metrics is introduced. It decimates the vertices and triangles of the model with little loss of rendering performance, and it enables the real-time rendering of the model.4. The visualization system based on OpenGL is implemented.5. A point extraction method, which shapes an object with a set of points, is introduced as an exploration on the 3D visualization technique and the possible association between surface and volume rendering techniques.