The 3D Motorcycle Complex for Structured Volume Decomposition

Hendrik Brückler,Ojaswi Gupta,Manish Mandad,Marcel Campen
DOI: https://doi.org/10.48550/arXiv.2112.05793
2021-12-11
Abstract:The so-called motorcycle graph has been employed in recent years for various purposes in the context of structured and aligned block decomposition of 2D shapes and 2-manifold surfaces. Applications are in the fields of surface parametrization, spline space construction, semi-structured quad mesh generation, or geometry data compression. We describe a generalization of this motorcycle graph concept to the three-dimensional volumetric setting. Through careful extensions aware of topological intricacies of this higher-dimensional setting, we are able to guarantee important block decomposition properties also in this case. We describe algorithms for the construction of this 3D motorcycle complex on the basis of either hexahedral meshes or seamless volumetric parametrizations. Its utility is illustrated on examples in hexahedral mesh generation and volumetric T-spline construction.
Graphics
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: how to decompose three - dimensional volumetric objects into a relatively small number of regularly - structured cuboid blocks, thereby simplifying the processing of complex geometric structures. Specifically, the authors propose the concept of 3D motorcycle complex, which is an extension of the 2D motorcycle graph in three - dimensional space. Through this method, a volumetric object with a suitable directional structure can be segmented into regular cuboid blocks without generating as many segmentation blocks as the traditional base complex. For example, given seamless parameterization or hexahedral meshes, the 3D motorcycle complex can significantly reduce the number of segmentation blocks. In the example given in the paper, the motorcycle complex only generates 79 blocks, accounting for 5.8% of the 1,352 blocks of the base complex. This simplification not only improves computational efficiency but also makes subsequent operations such as hexahedral mesh generation and volumetric T - spline construction simpler and more efficient. In addition, the paper also describes the algorithm for constructing the 3D motorcycle complex and shows its advantages in different application scenarios. The specific steps include: 1. **Expansion process**: Starting from singularities, gradually specify partial isoparametric surfaces as the boundaries of blocks. 2. **Ensure decomposition properties**: It may be necessary to specify some additional boundaries to ensure the required block decomposition characteristics. 3. **Shrink redundant boundaries**: Shrink boundaries in areas that are too aggressive during the initial expansion process to remove unnecessary complexity. Overall, this paper aims to provide a new, more efficient three - dimensional volume decomposition method to deal with complex problems in the fields of computer graphics and geometric processing.