Study On Virtual Endoscopy Of 3D Data Fields

Virtual endoscopy is a hot research topic in the field of scientific visualization, which unique technique is an important application of computer, computer graphics, image processing and virtual reality in endoscopy system. It is a advanced medical image post-processing system emerging only several years ago. Since it is a non-invaded, non-touched means and non-hurt examination method, virtual endoscopy is widely used in diagnostic and teaching in anatomy recently.Some key techniques and realizations of virtual endoscopy are systematically investigated in this thesis. Firstly, for the medical image segmentation problem, a fast fuzzy c-means (FCM) segmentation algorithm is proposed for the region-of-interest (ROI). Secondly, an interactive 2D histogram segmentation algorithm for interactively segmenting 3D data is also proposed. With this fast interactive segmentation method, scientists can study the interesting parts conveniently. Thirdly, the interactive virtual endoscopy techniques are investigated, and an automatics wandering fashion in virtual endoscopy system implemented. For the false contour problem of the surface rendering, an anisotropic diffusion smoothing algorithm is presented for surface blur. Finally, by analyzing the characteristic virtual endoscopy and volume data virtualization, an integrated volume virtual endoscopy algorithm is proposed to further improve the vivid effect of the virtual endoscopy system.