A Novel Double Periodic Conformal Flattening Algorithm for Genus-One Surfaces

Zhong-Heng Tan, Tiexiang Li, Wen-Wei Lin, Lok Ming Lui · 2024

Read A Novel Double Periodic Conformal Flattening Algorithm for Genus-One Surfaces by Zhong-Heng Tan, Tiexiang Li, Wen-Wei Lin, Lok Ming Lui (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-2412-19052.txt.

Introduction

Zhong-Heng Tan and co-authors present, in a 2024 math paper, a method to flatten donut-shaped surfaces onto a flat sheet while keeping angles true and edges matched in a doubled periodic way. Such flattening serves texture mapping and medical imaging. Their method seeks the map that least distorts angles, under matched-edge rules.

The search reduces to a clean quadratic problem with a sparse linear system at its heart. The best boundary shape and shift vectors come out together with the map. No cut path needs to be chosen, so no seam distortion appears. One-to-one mapping is assured under a mild weight rule, with a fix for general meshes. Tests ran five times faster than the best rival.

Catalog record

CREATORZhong-Heng Tan, Tiexiang Li, Wen-Wei Lin, Lok Ming Lui
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-19052

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Novel Double Periodic Conformal Flattening Algorithm for Genus-One Surfaces Zhong-Heng Tan, Tiexiang Li, Wen-Wei Lin, Lok Ming Lui Abstract In this paper, we propose novel parameterization methods for genus-one surfaces, called the Double Periodic Conformal Flattening (DPCF) algorithm. The desired conformal map is obtained by minimizing a conformal energy functional under periodic boundary conditions, which is characterized as an easily solvable quadratic functional minimization problem, yielding a sparse linear system. The proposed DPCF algorithm offers several key advantages: (a) the optimal boundary and periodic translation vectors are obtained simultaneously with the conformal map; (b) the resulting map is independent of the chosen cutting path, thus introducing no extra conformal distortion near the cut seams; (c) bijectivity is guaranteed under the positive edge weights condition, which can be satisfied by, e.g., using an intrinsic Delaunay triangulation. Based on this guaranteeing, a simple strategy is employed to ensure bijectivity of the resulting maps for general triangulations. Numerical experiments illustrate that DPCF algorithm exhibits high accuracy and a 5 fold improvement over the state-of-the-art algorithm in terms of efficiency. Applications on texture mapping and medical imaging illustrate the practicality of our developed algorithm. Key words— periodic conformal flattening, genus-one surfaces, cutting path independence, bijectivity 1 Introduction The advent of high-precision 3D technology has given rise to a plethora of surfaces that are characterized by high resolutions and complex geometries, which poses challenges to direct surface manipulation methods. Surface flattening is a fundamental technique that aims to flatten a surface onto a plane, transforming the task of processing a complicated surface into a simple 2D domain-related procedure. Generally, flattening maps should be bijective or locally injective to ensure optimal visual effects and high processing performance in practical applications, which are the focuses of some works [ 15 , 36 , 37 ] . The common goal of surface flattening is to reduce the degree of flattening distortion as much as possible. Conformal flattening is a flattening technique for minimizing conformal distortion, which preserves the local shape by the conformal map. Consequently, this approach plays a significant role in computer-aided engineering [ 27 , 5 ] and computer graphics [ 19 , 28 ] . For the genus-zero surface with boundary, various algorithms and methods have been developed to achieve conformal surface flattening, including most isometric parametrizations (MIPS) [ 21 ] , angle-based flattening [ 32 , 33 ] , LSCM [ 28 ] , DCP [ 10 ] , circle patterns [ 25 ] , spectral conformal parameterization [ 29 ] , quasi-conformal approach [ 9 , 7 , 6 ] and discrete conformal equivalence [ 16 ] . Ricci flow [ 24 , 38 ] is an effective method for computing conformal factors and subsequently constructing the conformal map. Similarly, the Yamabe equation [ 31 ] is also used to flatten surfaces with various boundary conditions. Conformal energy minimization has been proposed for disk parameterizations [ 40 , 26 ] and free boundary flattening [ 22 ] . Closed surfaces such as spheres and tori can be cut along a graph tree to be disk-topology, so that the cut surfaces are flattened to \mathbb{R}^{2} [ 30 , 2 , 34 , 12 ] . In this scenario, the cut seam correspondence should be taken into consideration. Generally, each pair pf boundaries from the same seam should have identical shape, i.e., they differ only from a rigid transformation. These maps are called seamless parameterization. Harmonic map based method [ 1 , 4 , 20 ] is a common approach to computing the seamless parameterization; that is, generate weights for each edge and solve a discrete harmonic equation with cut seam constraints. Floater [ 13 ] proved that the harmonic map is bijective if edge weights are positive and the flattened domain is convex. The positive edge weights could be satisfied by restricting intrinsic Delaunay mesh [ 1 ] or modifying the edge weights, such as mean value coordinates [ 14 ] . However, we can hardly obtain a convex domain under the seam constraints. For this problem, it is common to add bijective constriants [ 4 , 20 ] or removing non-bijective elements during iterating computation [ 35 , 37 ] to obtain locally injective maps or bijective maps. Furthermore, the boundary condition is not sufficient to guarantee the uniqueness of the solution. Then additional conditions and constraints, e.g., fixing rigid transformation [ 1 ] and some vertices [ 4 , 20 ] , are prescribed, which may lead to larger distortion. On the other hand, Gu et al. [ 18 , Chap. 11] and Yueh et al. [ 39 ] utilized the de Rham cohomology approach to compute conformal parameterizations for genus-one surfaces, respectively, which computes the associated holomorphic 1-forms first, and then obtains the corresponding holomorphic (conformal) maps by integrating along a suitable path from a fixed starting point to each vertex. From a theoretical point of view, this approach can be used to understand the relationship between holomorphic 1-forms and maps, and to produce a relatively accurate conformal map on the basis of rigorous theoretical guarantee. Gortler et al. [ 17 ] proved the bijectivity of holomorphic 1-form methods. However, from the numerical perspective of finding the conformal map, the computations of holomorphic 1-forms and subsequent line integrals on the de Rham cohomology theory [ 39 , 18 ] is truly a detour. In this paper, we aim to compute conformal seamless parameterizations for genus-one surfaces via conformal energy minimization and prove the bijectivity of our proposed algorithm. Since the flattening map produced with optimal translation correspondence for genus-one surfaces leads to a periodic tiling on \mathbb{R}^{2} , we call the proposed approach ‘double periodic conformal flattening’ (DPCF). The main contributions of this paper are threefold. • We propose an efficient and reliable numerical method, namely, DPCF for genus-one surfaces. Based on conformal energy minimization with identical cut seam constraints, the optimal boundary associated with translation vectors are obtained simultaneously with the map by only solving a sparse linear systems with two-column right hand side. The proposed algorithms have better performance in terms of accuracy and, in particular, provide efficiency improvements of approximately 5 times over the recently developed state-of-the-art algorithm [ 39 ] . • We show that the maps resulting from the developed algorithm are independent of the selected cutting path and that no additional conformal error is induced near the cutting path, which reflects the stability of our method. Based on this result, we further prove that the translation vectors only depend on the homology basis to which the cutting paths belong. • We generalize the bijective result in [ 13 ] and prove that the computed conformal map is bijective under the positive edge weights condition, e.g., the surface is an intrinsic Delaunay triangulation (IDT). We further develop a bijectivity guarantee strategy for non-Delaunay triangulations. Numerical experiments illustrate the significant effectiveness and accuracy improvement of the strategy. This paper is organized as follows. In Section 2 , we briefly introduce the concept of a discrete surface and review conformal energy functional. In Section 3 , we derive conformal energy minimization with periodic boundary conditions and propose DPCF algorithm. In Section 4 and 5 , we prove

Read the complete text