APPLICATION-DRIVEN QUANTITATIVE ASSESSMENT OF APPROACHES TO MESH GENERATION

In surgical simulation, it is common practice to use tetrahedral meshes as models for anatomy. A variety of mesh generators can automatically create tetrahedral meshes from segmented anatomical volumes, each offering its own set of unique attributes. To the authors knowledge, there has been little work directly assessing the suitability of different mesh generation schemes for surgical simulation. This paper seeks to address this issue by assessing the use of four mesh generators in a surgical simulation scenario, and examining how they affect simulation precision. The comparisons are performed against high-resolution reference meshes, in which we propose a novel error metric for assessing the efficacy of mesh generation schemes