Research On Virtual Endoscopy Path Planning

Virtual Endoscopy(VE) is the typical application of computer science in medical field.It utilizes medical images as original data to simulate traditional optical endoscopy by integrating technology of vitual reality, computer graphics, image processing techniques, the vitualization of scientific calculation. Since it is noninvasive, untouched and free of risks, it overcomes the shortcoming of the traditional optical endoscopy, which has to be inserted into patients'bodies. It also resolves the question that there is no access to the inside of certain organs.Some key techniques and realizations of VE are systematically investigated in this thesis. It consists of the reading of CT data, the pre-processing of medical images,the techniques of segment tissues, 3D Reconstruction and the techniques of path planning.Great stress is laid on 3D Reconstruction and the techniques of path planning in this thesis.The primary work of studies are as follows:First of all, Study the loading of the chest CT data and the pre-process of filtering and interpolation. The next , the semi-automatic thresholding algorithm based on thresholding segmentation is proposed.The CT images are segmented using the methods. In addition, Marching Cubes algorithms is improved in the process of VE rendering algorithm to improve the efficiency of rendering. Finally, based on the technique of distance transform method, this thesis puts forward a distance transform method on the basis of maximum-cost spanning tree, which is able to create central path quickly and to solve some branch questions.