Research On The Algorithms For Medical Visualization

Medical visualization, a multi-disciplinary subject, is now a hot spot of research on image processing. It is an important application of computer graphic and image processing to biomedical domain. In recent years, the medical visualization techniques for medical images are widely used in diagnosis, surgery planning and simulating, plastic and artificial limb surgery, radiotherapy planning and teaching in anatomy. Nevertheless, visualization is not limited to the medical field and has great potentials in nondestructive flaw detection, geology and aerography domains as well.In this thesis, some key issues of medical visualization techniques are deeply and systematically investigated. First, filter, interpolation and some other two dimensional pre-processing techniques of medical image slices are reviewed. A real-time bilinear interpolation method is proposed which can improve the performance enormously. Then, on the bases of those, the study of direct volume rendering is carried out. A directional filter kernel based splatting method is presented which can render an image more dedicatedly than the traditional one does thus generating an image of higher quality, especially when the object is zoomed or put close to the view point. Finally, a comprehensive analysis of the hardware accelerated texture mapping renderings is covered, and a texture splatting method is also implemented, which make it possible to be widely applied to industry level. A plenty of medical and industrial testing dataset illustrate the effectiveness of the algorithms presented and the applications in various fields.