The Reserach Of Fault Medical Image's Three Dimension Reconstruction

The three dimension reconstruction of fault medical image is a combination of computer graphic image, image processing and visualization technology , and so on. It is a hot research widely used in medical. It can reconstruct two dimension fault medical image to three dimension realistic medical image. And it can also help the diagnosis of a doctor with less experience because it directly gives the organs and organizations'shape. The three dimension reconstruction image help doctors enhance the diagnostic efficiency. This Algorithm has a big significance in medical research and clinical.This article mainly studies surface rendering and volume rendering, including marching cubes algorithm and ray casting algorithm. First, it introduces the principle of the three dimension reconstruction and analyses the key technology. Second, it points out the deficiencies of this algorithm. Third, it improves this algorithm and analyses the improved effect.In surface rendering, we process the fault image firstly. Then, we get the body data which can reconstruct right now. Next, we introduce splicing contour line algorithm and marching cubes algorithm in the surface rendering. In marching cubes algorithm, we adopt the simplified middle point interpolating method in order to improve the calculation efficiency in traditional algorithm. According to the ambiguity on and in the cube, we study the strategy to eliminate this ambiguity. Because it gives lots of triangle patches, it can not display in time. So, we need reduce triangle patches through vertex merging algorithm. Finally, we smooth the triangle patches so that we can observe better. Through experiment contrast, the improved algorithm proves the feasibility.In volume rendering, we firstly introduce the light model . This article adopts light absorbing &emission model. Secondly, we introduce the traditional ray casting algorithm's principle and process. Then, we analyse the key technology including data classification, color and opaque assignment, again sampling and image composite. We point out the defect that it emits so many light and has so many sampling points in a light that it wastes lots of time and space resource. In the improved algorithm, we cut the scale of three dimension data field so that reduce the empty sample points. By establishing polygon scan and starting grid, we reduce the sample points in a light. Besides, the light can be reduced through adaptive casting, too. Finally, according to the experiment contrast, the improved algorithm is feasible.In the end, this article introduces kinds of visual tools and clarifies their function feature.