How to Flatten a Soccer Ball

Kaie Kubjas, Pablo A. Parrilo, Bernd SturmfelsSpringer INdAM Series 2017 · DOI · arXiv · OpenAlex

Abstract

This paper presents a rigorous case study in real algebraic geometry, focusing on the computation and characterization of 2D images of 3D semialgebraic sets—specifically a "soccer ball" or unit ball—under polynomial mappings. We analyze the highly singular curves that define the boundaries of these flattened shapes and demonstrate how complex topological features, such as multiple holes, can arise from relatively simple polynomial projections.

Introduction: The Geometry of Projections

In diverse fields such as computer vision, robotics, and optimization, understanding the "shadow" or projection of a high-dimensional object onto a lower-dimensional space is a fundamental challenge. Mathematically, this involves mapping a semialgebraic set (a shape defined by polynomial inequalities) into a plane using a polynomial function. While Tarski’s theorem on quantifier elimination ensures that the resulting 2D image is also semialgebraic, obtaining a practical, quantifier-free description is computationally intensive. This study addresses the exact algebraic boundary of these flattened shapes and their topological complexity.

Methodology and Mathematical Framework

The authors utilize a combination of classical algebraic geometry, tropical geometry, and computational tools to dissect the projection process.

The 2D image of a 3D ball under a polynomial map, showing the complex boundary curves and interior point distribution.

Analytical Tools

  • Branch Locus Analysis: The boundary of the flattened image is shown to be composed of two distinct curves: the branch locus of the map itself ($p$) and the branch locus of the map restricted to the surface of the ball ($q$).
  • Tropical Geometry: This is employed to determine the Newton polygons of the boundary curves, facilitating predictions regarding their degree and algebraic complexity.
  • Cylindrical Algebraic Decomposition (CAD): While CAD is the standard algorithm for quantifier elimination, the authors provide "human-readable" versions to offer deeper intuition into the process.
  • Chebyshev Polynomials & Lissajous Curves: These are used to construct specific maps that "fold" the ball into shapes with a predictable number of internal holes.

Experimental Setup

The study relies on synthetic mathematical benchmarks rather than traditional datasets. These include the unit ball ($u^2 + v^2 + w^2 \le 1$), symmetric maps using elementary symmetric polynomials, and custom-built Lissajous maps designed to test topological limits.

Key Insights: Singularities and Topology

A primary finding is that the edge of a flattened ball is not a simple smooth line but is defined by the product of two polynomials, $p$ and $q$. While $p$ is a property of the map, $q$ depends on both the map and the geometry of the original ball.

For generic polynomial maps, these boundary curves are "highly singular," characterized by numerous self-intersections (nodes) and cusps. Furthermore, the authors demonstrate that a solid convex ball can be "flattened" into a shape with holes. Using Chebyshev polynomials of degrees $d_1$ and $d_2$, it is proven that one can create a flattened ball with exactly $\frac{(d_1-1)(d_2-1)}{2}$ holes. This implies that convexity is rarely preserved under polynomial mapping; most projections result in non-convex "splats."

Quantitative Results

The paper provides explicit formulas for the degree, genus, and singularity count of the boundary curves.

Complexity metrics for the boundary curves V(p) and V(q) across various polynomial degrees.

As shown in the table above, the complexity grows rapidly with the degree of the mapping. For instance, a degree-4 map results in a boundary curve $V(q)$ with a degree of 56 and 1,449 singular points. The authors also establish a lower bound for the number of holes in the complement of the flattened image using Lissajous-style maps and present a 5-step algorithm for computing quantifier-free representations that is more interpretable than standard software outputs.

Limitations and Future Directions

Despite the theoretical advances, significant computational hurdles remain. The underlying CAD process is doubly exponential, making scaling to higher dimensions or degrees difficult. Additionally, while the paper demonstrates mapping a ball onto specific shapes like hexagons, the general problem of constructing polynomial maps to transform a ball into any arbitrary convex polygon remains an open challenge. Future work may also address the gap between generic cases and the "special" polynomials often encountered in engineering applications.

Summary updated

BibTeX
@inproceedings{kubjas2017how,
  author = {Kaie Kubjas and Pablo A. Parrilo and Bernd Sturmfels},
  title = {{How to Flatten a Soccer Ball}},
  booktitle = {Springer INdAM Series},
  year = {2017},
  doi = {10.1007/978-3-319-61943-9\_9},
}

Loading related content...

Figure