Abstract
The covariogram $g_K$ of a convex body $K$ in $\mathbb{E}^d$ is the function which associates to each $x$ in $\mathbb{E}^d$ the volume of the intersection of $K$ with $K+x$. In 1986, G. Matheron conjectured that for $d=2$, the covariogram $g_K$ determines $K$ within the class of all planar convex bodies, up to translations and reflections in a point. This problem is equivalent to challenges in stochastic geometry, probability, and the phase retrieval problem in Fourier analysis. It is also relevant for determining the atomic structure of quasicrystals from X-ray diffraction. In this paper, we confirm Matheron's conjecture completely.
Introduction
The covariogram of a set $K$, denoted $g_K(x)$, measures the area of overlap between $K$ and its translated version $K+x$. This geometric "fingerprint" is central to several fields. In crystallography, it relates to the Patterson function used to solve crystal structures. In Fourier analysis, it corresponds to the phase retrieval problem: reconstructing a shape's characteristic function from the magnitude of its Fourier transform.
For over twenty years, the mathematical community sought to prove Matheron's 1986 conjecture: that in two dimensions, this overlap data uniquely identifies a planar convex body (up to translation and point reflection). While proven for specific cases like polygons or smooth bodies with positive curvature, a general proof for all planar convex bodies remained elusive until now.
Methods and Theoretical Framework
The proof strategy developed by Averkov and Bianchi moves beyond standard geometric approaches by analyzing the analytical properties of the covariogram function, specifically its second-order partial derivatives.
Inscribed Parallelograms
The authors utilize the fact that for any vector $x$ in the support of the covariogram, there exists a corresponding parallelogram inscribed within the convex body $K$. The vertices of this parallelogram lie on the boundary of $K$, and the way these vertices shift as $x$ varies allows for the reconstruction of the boundary.
The Hessian-Geometry Link
A critical component of the proof is the relationship between the Hessian matrix $G(x)$ of the covariogram and the local geometry of the body $K$. The authors demonstrate that the second derivatives of $g_K(x)$ encode information about the curvature and orientation of the boundary at specific points.
By synthesizing previous results for polygons and smooth bodies, the authors provide the necessary logical bridge to cover all planar convex bodies, regardless of their boundary's smoothness or the presence of flat segments.
Key Results
The central result of this work is the complete confirmation of Matheron's conjecture in two dimensions:
Theorem: Every planar convex body is uniquely determined within the class of all planar convex bodies by its covariogram, up to translations and reflections.
This result establishes the covariogram as a complete descriptor for 2D convex shapes. Notably, this uniqueness is dimension-specific; the conjecture is known to be false in dimensions $d \ge 4$, where distinct shapes can share identical covariograms.
Conclusion and Future Directions
The resolution of the 2D case marks a significant milestone in geometric tomography. However, the problem remains open in three dimensions for general smooth convex bodies, although it has been solved for 3D polytopes. Furthermore, the requirement of convexity is essential, as non-convex sets are known to be non-uniquely determined by their covariograms. Future research will likely focus on extending these results to 3D and identifying the minimal additional information required for uniqueness in higher dimensions.
Summary updated