Measuring Errors for Huge Semi-regular Meshes
Résumé
The softwares Mesh and Metro are widely used for measuring geometrical differences between two surfaces. But those softwares cannot be used to compute the surface-to-surface distance for huge semi-regular meshes because of the memory capacity. So, estimating the quality of remeshing or geometry compression algorithms cannot be done for such data. To overcome this problem, we propose an original algorithm for computing the surface-to-surface distance even for huge semi-regular meshes. The method consists in exploiting the relevant multi-level structure of a semi-regular mesh for loading successively small regions of it and computing the symmetrical distance between them and the irregular mesh. Experimentally, the results obtained with the proposed method are similar to those obtained with MESH, while using a smaller memory size. This latter can reach only 2% of the size of the semi-regular mesh. Then, we show that our approach allows to compute the surface-to-surface distance for huge semi-regular meshes.