A parsimonious approach to C 2 C^{2} cubic splines on arbitrary triangulations: Reduced macro-elements on the cubic Wang–Shi split
Tom Lyche, Carla Manni, Hendrik Speleers · 2024
Read A parsimonious approach to C 2 C^{2} cubic splines on arbitrary triangulations: Reduced macro-elements on the cubic Wang–Shi split by Tom Lyche, Carla Manni, Hendrik Speleers (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-18323.txt.
Introduction
Tom Lyche and two co-authors build leaner spline tools for smooth shapes over triangle meshes, in a 2024 math paper. Splines are patchwork curves used to model smooth surfaces in design and simulation. The full cubic spline space on a split triangle mesh is rich but heavy. The team carve out subspaces that keep cubic power while using far fewer control values. Some versions need data only at mesh corners and edges, and one needs corners alone.
The leanest space uses just six values per mesh corner. Each patch can be built on its own from local data, in a macro-element style. A local basis handles the tricky split geometry behind the scenes. The authors say the savings suit large models where each degree of freedom costs time.
Catalog record
| CREATOR | Tom Lyche, Carla Manni, Hendrik Speleers |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 57 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-18323
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A parsimonious approach to C 2 C^{2} cubic splines on arbitrary triangulations: Reduced macro-elements on the cubic Wang–Shi split Tom Lyche, Carla Manni, Hendrik Speleers Abstract We present a general method to obtain interesting subspaces of the C 2 C^{2} cubic spline space defined on the cubic Wang–Shi refinement of a given arbitrary triangulation 𝒯 {\mathcal{T}} . These subspaces are characterized by specific Hermite degrees of freedom associated with only the vertices and edges of 𝒯 {\mathcal{T}} , or even only the vertices of 𝒯 {\mathcal{T}} . Each subspace still contains cubic polynomials while saving a consistent number of degrees of freedom compared with the full space. The dimension of the considered subspaces can be as small as six times the number of vertices of 𝒯 {\mathcal{T}} . The method fits in the setting of macro-elements: any function of such a subspace can be constructed on each triangle of 𝒯 {\mathcal{T}} separately by specifying the necessary Hermite degrees of freedom. The explicit local representation in terms of a local simplex spline basis is also provided. This simplex spline basis intrinsically takes care of the complex geometry of the Wang–Shi split, making it transparent to the user. Keywords: B-splines, Simplex splines, Macro-elements, Triangulations 1 Introduction Smooth splines on triangulations are a popular tool to approximate or interpolate function data [ 3 ] . Following the classical finite element approach [ 2 ] , such spline spaces are often characterized by means of Hermite degrees of freedom associated with the vertices, edges, and/or triangles of the triangulation. In practical situations, however, data values are not always available in the interior of the triangles and are often only attached to the vertices of a given triangulation. Therefore, it is preferable to avoid Hermite degrees of freedom associated with the triangles, and even better, to have them only associated with the vertices. Additionally, for the sake of approximation quality, the provided construction should ensure reproduction of polynomials up to the highest possible degree. Finally, local constructions that can be performed on each triangle separately are highly attractive from a computational viewpoint. In classical finite element constructions, it is a common procedure to remove Hermite degrees of freedom associated with the triangles or edges by expressing them as a linear combination, with suitable coefficients, of the degrees of freedom associated with the vertices [ 2 ] . However, the full approximation order is not always preserved in those constructions; this is, e.g., the case for the Bell element [ 1 ] . Expressing degrees of freedom in terms of other degrees of freedom can also be helpful in the construction of B-spline-like basis functions for reduced spline spaces on triangulations [ 13 , 14 ] . In the univariate case, splines of maximal smoothness have proven to be very useful. They maintain the same approximation order as less smooth splines of the same degree but involve fewer degrees of freedom and have less tendency to oscillate [ 10 , 11 ] . When moving to the bivariate setting on triangulations, maximal smoothness is still very appealing but becomes an arduous task to achieve [ 3 , 12 ] . Bivariate spline spaces with too low degree compared to the smoothness are exposed to several shortcomings: they may lack a stable dimension, optimal approximation power, and stable locally supported bases. To construct C^{2} cubic spline functions on a given triangulation {\mathcal{T}} , it is opportune to refine each triangle of {\mathcal{T}} by the Wang–Shi (WS) split of order three; see [ 4 , 16 , 15 ] . Even though the split is rather complex, the corresponding C^{2} cubic spline space possesses a stable dimension and can be simply characterized by means of certain Hermite degrees of freedom, of which there are six per vertex, three per edge, and one per triangle in the triangulation. These degrees of freedom allow us to locally build a spline function on each triangle of {\mathcal{T}} separately. Furthermore, the space is equipped with a local simplex spline basis [ 4 ] , which enjoys properties similar to the Bernstein polynomial basis on a triangle. Thanks to the characteristics of this local simplex spline basis, one can avoid to consider separate polynomial representations on each of the polygonal subregions of the WS split. Instead, there is a single control net to facilitate control and early visualization of a spline function over each triangle of {\mathcal{T}} . This makes that the complex geometry of the WS split is transparent to the user. A similar simplex spline construction for C^{3} quartic splines has been developed in [ 5 ] . In this paper, we present a general method to obtain interesting subspaces of the C^{2} cubic spline space defined on a WS refined triangulation. These subspaces are characterized by specific Hermite degrees of freedom associated with only the vertices and edges of {\mathcal{T}} , or even only the vertices of {\mathcal{T}} . Each subspace contains cubic polynomials, a necessary condition for attaining full approximation order. Moreover, any element of such a subspace can be constructed on each triangle of {\mathcal{T}} separately by specifying the necessary Hermite degrees of freedom. We also provide the explicit local representation of such an element in terms of the local simplex spline basis. Therefore, we provide a construction of C^{2} cubic splines on triangulations that is parsimonious in the sense that it produces spaces still containing cubic polynomials but requiring fewer degrees of freedom. The dimension of the considered subspaces can be as small as six times the number of vertices of the given triangulation. The remainder of the paper is divided into four sections. Section 2 collects some introductory material about the local C^{2} cubic spline space on a single WS refined triangle and the global C^{2} cubic spline space on a WS refined triangulation. Section 3 describes a procedure how to peel away Hermite degrees of freedom from the local spline space on each triangle and how to represent the resulting subspaces in terms of a local simplex spline basis. These local subspaces can then be assembled to form a global subspace containing cubic polynomials as outlined in Section 4 . Section 5 concludes with some final remarks. Since this paper is a follow-up to [ 4 ] , we use the same notation and terminology as introduced in [ 4 ] for ease of comprehension. 2 The cubic spaces {\mathbb{S}}_{3}^{2}(\Delta_{{{\rm WS}_{3}}}) and {\mathbb{S}}_{3}^{2}({\mathcal{T}}_{{{\rm WS}_{3}}}) In this section, we provide some preliminary material necessary for the rest of the paper. We first summarize the construction of C^{2} cubic splines on the WS split of order three of a given triangle by specifying suitable Hermite degrees of freedom [ 4 , 15 ] , and then extend it to obtain C^{2} cubic splines on an arbitrary triangulation {\mathcal{T}} . Subsequently, we recall from [ 4 ] the definition and main properties of a local simplex spline basis for the considered local space on any triangle of {\mathcal{T}} , and we describe its relation with the above mentioned Hermite degrees of freedom. Figure 1: {{\rm WS}_{d}} splits for d=2,3,4 . 2.1 Definition of the spaces Let {\boldsymbol{p}}_{1},{\boldsymbol{p}}_{2},{\boldsymbol{p}}_{3} be three noncollinear points in {\mathbb{R}}^{2} and consider the triangle \Delta:=\langle{\boldsymbol{p}}_{1},{\boldsymbol{p}}_{2},{\boldsymbol{p}}_{3}\rangle