Research And Realization Of Skeleton Extracting Technique For 3D Modeling

Prevailing in the world of Computer-Aided Design, digital museum, medical imaging, scientific visualization, virtual reality, computer graphics and gaming environment, Three-dimensional (3D) data become more and more a common feature of nowadays multimedia. While the 3D representation is invaluable, many applications require alternate"compact"representations of these models, in order to satisfy various requirements related to storage space or computational overhead at application run time. The line-like representation is a one-dimensional (1D) abstraction of a three-dimensional (3D) object, consisting of a set of curves embedded in 3D space. This line-like representation of a 3D object is also known as the centerline or the curve-skeleton. As a reduced representation, the curve-skeleton capture the essential shape (topology and geometry) of the underlying 3D object in an easy to understand and very compact form, also increase the efficiency of memory usefulness and compress ratio.In the past twenty years, a large numbers of curve-skeletonization methods were developed in 3D models. However there isn't a uniform experimental tool for curve-skeletonization until now. The major contribute of this thesis is that developing a visual, open and extensible experimental tool with the various existing curve-skeletonization methods. Firstly, the definitions of skeleton are discussed in detail, and the popular skeleton algorithms are reviewed. Through comparing, the advantage and disadvantage of each class were proposed. Secondly, an improved voxelization algorithm from mesh data to volume data is proposed after the advantage and disadvantage of existing algorithms were discussed. By taking an adaptive minimal bounding box as computational unit and Euclidean distance as measurement criterion, polygonal meshes can be voxelized fast and implemented on PC instead of workstation. The performance of implementation was reported comparing with original method. Finally, the implementation of the experimental tool was discussed in detail. One algorithm from each class has been implemented and tested on the same set of 3D objects. With the results we discussed to what extent each methodology achieves the different curve-skeleton properties.