Research On MRI 2D Images Processing And 3D Visualization

The dissertation is an important part of the "the tenth five years" key subject building project of NUAA, which is called The Computer Integration Diagnoses and Therapy System. The study solves the key problems of 2D image pre-processing in the project, and it is ready for the technology of 3D visualization. So the dissertation studies the 2D image pre-processing, such as image enhancement, edge detection and image segment, then discusses the technology of 3D visualization based on the better pre-processing in a way. Now, magnetic resonance imaging system can offer to doctors the 2D slice image information of tissue and organs of the human body, but the slice images are not easy to get relative space position relation between every organ. Yet, 3D visualization technology can show organs and pathological position of the human body into three-dimensional form, which offers visual information for clinical medicine. However, the better 3D visual results depend on the effect of 2D image pre-processing. So study on the technology of MRI images pre-processing and 3D visualization, has important significance on science and worthiness in practical application. The main research contents of MRI image pre-processing include image enhancement, edge extraction and image segmentation. Because the noise of medical images is larger, conventional MRI image enhancement arithmetic will enhance noise signals of the original image. By improving image enhancement arithmetic based on wavelet transform, we can improve the contrast of images and resolve the problem of the noise magnified. And MRI images have plenty of weak edge characteristic of symptom information; the traditional methods of edge extraction often extract noise as edges. So, based on wavelet transform, we presented a new feasible edge extraction arithmetic, which experiments indicate that the edge detection is effective. Then we use the segmentation method based on edges to segment the MRI images. The image pre-processing is effective to improve quality of MRI images. Finally, we study the improvement of 3D visualization's speedup based on the better 3D visual results and use Ray-casting arithmetic to reconstruct 3D model faster. The improved arithmetic based on wavelet transform in the dissertation can get the better effects. And they are useful to improve 3D visualization. The other rule of 3D visualization is the speed, so the elementary study of the speedup method is put forward on the latter part and is ready for the future processing.