Research About 3D Model Simplification Methods

As the computer graphics and processing capability of computers improves, 3D model has become the forth type of media, which widely used in virtual reality, computer animation & games, manufacture and other areas. While the accuracy of the 3D models grows with the development of the 3D scanning and modeling techniques, the volume of their data sets grows even rapidly, which brings great pressure to the rendering, transmitting and editing system of computer. One way to solve this problem is to simplify these complex 3D models and to perform multi-resolution modeling. In 3D model simplification problem, researchers focus mainly on triangular mesh as it is well suited for describing 3D model. Polygonal mesh is a set of polygons or faces, which construct the model's surface. Especially if a polygonal mesh is comprised of triangles, it is called triangular mesh.A lot of 3D mesh simplification methods have been proposed, for example: Lounsbery et al. proposed a wavelet-based method; Schroeder et al. proposed an algorithm based on vertex decimation; Rossignac et al. proposed an algorithm based on vertex clustering; Hoppe et al. proposed an algorithm based on interactive edge collapse. In recent years, some methods dealing with the surface properties of 3D meshes have been described: Okuda proposed a simplification algorithm not only considered the surface geometry but also the texture property; Fahn proposed a new scheme based on the quadric error metric to preserve face colors and bounding edges during the simplification process.It is easy to find that there are two problems existing in the 3D mesh simplification methods proposed above: (1) in most precious works, surface geometry, color, and texture are considered except closeness property; (2) in some applications of 3D meshes, user is only interested in one part of a 3D mesh, and hope it could preserve a high resolution, while the rest parts could meet the basic requirement with low resolution. However, the recent simplification methods treat the same with the interested and uninterested region, which cannot satisfy the visual requirement that the interested region preserve a high resolution.In this paper, the author proposes a simple and efficient 3D mesh simplification method with closeness which could successfully preserve closeness of the original mesh during the simplification process, and a 3D mesh simplification method based on ROI which adapts to the situation that the user is only interested in one part of a 3D mesh.The main contribution of this paper includes: (1)Design a 3D mesh simplification algorithm based on Degree Centrality. It uses Degree Centrality Theory to compute a weight for every edge in the mesh, and the edge with smaller weight is always removed earlier. Compared with the precious methods, this one works very well in preserving both the appearance and closeness of the original mesh. The experimental results support the collusion obviously.(2)Propose a 3D mesh simplification algorithm based on ROI. It uses ROI coding technology to divide the original 3D mesh into two parts, the interested region and uninterested region according to the point of interest that is decided by the user. During the simplification process, edge-collapse operations are largely abandoned to the edges belonged to the interested region, which makes the interested region preserve a high resolution. Experiments results have proved that during the simplification this algorithm could successfully preserve a high resolution for the 3D part that the user is interested in while the rest parts could meet the basic requirement with lower resolution, which effectively save the storage space.