Detection of hierarchical intrinsic symmetry structure in 3D models

2017 Elsevier Ltd The structure of a 3D model is essential for identifying and understanding of the object. A large number of models show hierarchical structures, which are not only efficient for shape characterization, but also applicable to structural/topological editing. However, how to construct the hierarchical structure is challenging. Most of the existing methods construct the hierarchical structure of a model based on the local relationships among parts. We propose Hierarchical Intrinsic Symmetry Structure (HISS) to represent the structure of models and an automatic construction approach. The analysis of intrinsic symmetry is based on the skeleton of the model. First, we extract skeletons of semantic parts of a model and construct skeleton graph. Next, we calculate the similarities among parts. Then, we employ a voting strategy to determine the hierarchical structure centers according to the intrinsic symmetry among similar part pairs. Finally, we construct the HISS based on the skeleton graph and the detected hierarchical centers. We evaluate the representation and the algorithm with a set of experiments and applications, including model set analysis, model repairing and model synthesis.

Detection of hierarchical intrinsic symmetry structure in 3D models Academic Article