Title: Mesh Processing and Generation
Speaker:
Jun Wang, University of Wisconsin-Milwaukee
Date/Time: Tuesday, April 6th, 2010, 9:00 am
Location: CSRI Building/Room 95 (Sandia NM)
Brief Abstract:
- Surface mesh denoising
We propose a robust surface mesh denoising method that can effectively remove mesh noise while faithfully preserving sharp features. This method utilizes surface fitting and projection techniques. Sharp features are preserved in the surface fitting algorithm by considering an anisotropic neighborhood of each vertex detected by the normal-weighted distance. In addition, to handle the mesh with a high level of noise, we perform a pre-filtering of surface normals prior to the neighborhood searching. A number of experimental results and comparisons are given to demonstrate the excellent performance of this method.
- Surface mesh optimization
The smoothness and angle quality of a surface mesh are two important measurements of the goodness” of the mesh for downstream applications such as visualization and numerical simulation. We present a novel surface mesh smoothing method not only to reduce the mesh bumpiness, but to improve the angle quality as well. Our approach is based on the local surface fitting around each vertex using the least square minimization technique. The new position of the vertex is obtained by finding the Maximum Inscribed Circle (MIC) of the surrounding polygon and projecting the circle’s center onto the analytically fitted surface. The procedure above repeats until the maximal vertex displacement is less than a pre-defined threshold. The mesh smoothness is improved by a combined idea of surface fitting and projection, while the angle quality is achieved by utilizing the MIC-based projection scheme. Results on a variety of biological and engineering models are shown to demonstrate the effectiveness of this method.
- Volumetric mesh generation
We propose a new tetrahedral mesh generation algorithm to extract adaptive and quality volumetric mesh directly from arbitrary surface mesh model. A top-down octree subdivision for a surface mesh model is performed adaptively according to a feature sensitive geometric metrics. After tilling the space of octree with body center cubic lattices, a candidate tetrahedral mesh is generated without introducing any hanging nodes. Then, we use a modified marching tetrahedra method to extract the high quality tetrahedral mesh around the boundaries of model. By constraining the max depth-disparity of adjacent octree nodes, the angle quality of the tetrahedral mesh generated can be guaranteed with a specific value. The resulting mesh is well suited for subsequent applications, such as simulation, deformation and visualization.
CSRI POC: Pat Knupp, (505) 284-4565 |