Research On Techniques Of 3D Reconstruction Of Human Vessels

The visualization of the biological structure and function is essential to the understanding of health and disease of human being. In recent years, the integration of the information technology and medicine has a major impact on the level of medical education, the development of clinic medicine and the medical research. There are several research projects which are being carried out or will be carried out in this field, such as the visible human project (VHP) and the virtual human project. In November 2001, Chinese scientists conceived the idea of constructing the Chinese digital human at the 174th workshop of the Xiangshan Science Conference.The main research work of the thesis is 3D reconstruction of vessels as required under the background of the project of the digital virtual human. It includes three parts:The first part introduces the basic concept and the development of volume visualization and volume graphics firstly; then discusses and compares some principle algorithms in this field, such as the ray-casting algorithm, splatting algorithm, marching cubes algorithm, and so on.The second part analyzes the relative work in the field of 3D reconstruction of the vessels. According to the characteristics of the human vessels, a processing approach of 3D reconstruction of vessels from the CT or other 2D images is presented. In this approach, a special vessels'physical model is used, which can avoid the disturbance of the other organs in the human body and reduce the noise produced in the data acquisition stage. This approach based on the marching cubes algorithm can create the 3D surface model of the vessels and obtain the geometric information of the surface point. The experimental result shows the quality of resulted surfaces of vessels is satisfactory.The vessels have normally plenty of branches and complex topologic structure, a more concise and precise model than the pure and simple surface model is needed. In this thesis, the author uses the skeleton and centerline as the descriptor. After systematically introducing the concept of volume graphics, the third part discusses several skeleton extraction techniques and realizes an algorithm which can extract the vessels'skeleton and centerline. Every skeleton point has the distance information which indicates the distance from the point to the boundary.