This paper introduces a three-dimensional mesh generation algorithm for domains bounded by smooth surfaces. The algorithm combines a Delaunay-based surface mesher with a Ruppert-like volume mesher, to get a greedy algorithm that samples the interior and the boundary of the domain at once. The algorithm constructs provably-good meshes, it gives control on the size of the mesh elements through a user-de ned sizing eld, and it guarantees the accuracy of the approximation of the domain boundary. A noticeable feature is that the domain boundary has to be known only through an oracle that can tell whether a given point lies inside the object and whether a given line segment intersects the boundary. This makes the algorithm generic enough to be applied to a wide variety of objects, ranging from domains de ned by implicit surfaces to domains de ned by level-sets in 3D grey-scaled images or by point-set surfaces.

This paper proposes an algorithm to mesh 3D domains bounded by piecewise smooth surfaces. The algorithm handle multivolume domains defined by surfaces that may be non connected or non manifold. The boundary and subdivision surfaces are assumed to be described by a complex formed by surface patches stitched together along curve segments. The meshing algorithm is a Delaunay refinement and it uses the notion of restricted Delaunay triangulation to approximate the input curve segments and surface patches. The algorithm yields a mesh with good quality tetrahedra and offers a user control on the size of the tetrahedra. The vertices in the final mesh have a restricted Delaunay triangulation to any input feature which is a homeomorphic and accurate approximation of this feature. The algorithm also provides guarantee on the size and shape of the facets approximating the input surface patches. In its current state the algorithm suffers from a severe angular restriction on input constraints. It basically assumes that two linear subspaces that are tangent to non incident and non disjoint input features on a common point form an angle measuring at least 90 degrees.

Fig. 1. Surface meshes. Left: Meshing the boundaries of brain tissues with a tissue-dependent resolution. We requested the cortical surface (colored in dark gray) to have a resolution twice higher than other tissues. As a result, output surface meshes of other anatomical structures have a coarse resolution except in local regions where they neighbor gray matter. This is apparent in the magnified view of the white matter mesh (in light gray), in which the gray matter interface has been partially removed for visualization purposes. Right: The quality of the obtained surface meshes is much higher than with the marching cubes algorithm.

No abstract