Growing Cell Structures Learning a Progressive Mesh During Surface Reconstruction - A Top-Down Approach

Abstract

Growing Cell Structures (GCS) have been proven to be suitable for surface reconstruction from unstructured point clouds. The reconstructed triangle mesh can be represented compactly as a progressive mesh with integrated level of detail by storing only vertex split operations. However, half-edge collapse operations are used for GCS. In this paper, we present an improvement to a GCS-based surface reconstruction technique by converting a halfedge collapse to a more general vertex removal to create a progressive mesh. We have evaluated the new technique with respect to running time overhead and mesh quality. Results indicate that this technique can be used for efficient surface reconstruction. We will use the presented findings as basis for future research.