Teeth And Jaw Three-dimensional Point Cloud Data Visualization Research

In the area of medically assisted diagnosis, Oral three-dimensional visualization isa very important method in clinical diagnosis. It can clearly show the shape of thepathological changes on the dental, and provide more convenience and a clear way todiagnose. So it has gradually become an important research direction and hotspot in thearea of medical computer applications. In this disertation, the main content isthree-dimensional visualization of the dental point cloud data. In view of the realapplication requirements, this dissertation propose a series of feasible scheme aboutpoint cloud processing about de-noising, simplify and reconstruction.The dissertation firstly analyzes the characteristics of the dental point cloud data,and then two different methods, according to the requirements of practical application,are present to noise reduction, where grid method is selected to reducing odd noise andbilateral filtering is used to fair the point cloud data. Finally, grid method is used tosimplify the data because of the characteristic of the point cloud density is too high. Andthe de-noising processing which does not change the precision of the model whilegreatly reudced the amount of point cloud data is the foundation of later high-speedprocessing. All these methods are carried out in experiment platform, and the resultsprove that these methods are viable, efficient and of high accuracy.In this dissertation, the problem of reconstruction of the dental point cloud data isanalyzed, and various point cloud reconstruction algorithms are compared. To reach therequirements of accuracy and time for the practical application, a fast and exact surfacereconstruction algorithm based on Ball Pivoting is presented. Finally, this method iscompared with other methods in view of precision, complexity, and stability. The resultshows that the Ball Pivoting Algorithm is time efficient, high accurate, and stable, andthe quality of the output mesh is better than existing techniques.As a consequence through the work, a prototype system of3D dental point clouddata processing and rendering has been developed by using VC++on Windows XPplatform and the third-party library like wxWidgets, Opengl and Vcglib. Users can usethe system to perform the function such as de-noising, reconstruction, rendering the3D dental model, and can achieve a variety of format point cloud data input and output inthis system. At last, the system is tested by using several large scanned datasets, and theresults show that the system is stable and accurate to meet the requirement.