Efficient Computation of 3D Skeletons by Extreme Vertex Encoding

Abstract

Abstract. Many skeletonisation algorithms for discrete volumes have been proposed. Despite its simplicity, the one given here still has many theoretically favorable properties. Actually, it provides a connected surface skeleton that allows shapes to be reconstructed with bounded error. It is based on the application of directional erosions, while retaining those voxels that introduce disconnections. This strategy is proved to be specially well-suited for extreme vertex encoded volumes, leading to a fast thinning algorithm.

Virtual Porosimeter

“…The algorithm performs the following process for each reachable connected component. First, it computes its skeleton, using the method presented in [8] and [1], which obtains a surface (2D) skeleton with the same connectivity as the original model. Then, it identifies those skeleton voxels that are connected with the exterior and initializes a list (globalseed list) with them.…”

An extended concept of voxel neighborhoods for correct thinning in mesh segmentation

“…During the thinning process, it must be ensured that the resulting voxel set is topologically equivalent to the original object. To guarantee this, we employ the criterion proposed by Bertrand and Malandain in [8] for classifying which voxels may be thinned within iteration. In order to achieve adequate smoothness of the voxel skeleton for our purposes, we supplement our thinning method with an additional criterion, which was proposed by Tsao and Fu in [11] and ensures uniform thinning.…”