Some inequalities for the L_p-curvature images

Bin Chen, Weidong WangJournal of Mathematical Inequalities 2019 · DOI

Abstract

Lutwak introduced the notion of $L_p$-curvature image and proved an inequality for volumes of a convex body and its $L_p$-curvature image. In this article, based on the $L_p$-affine surface area and $L_p$-dual affine surface area, we establish the affine isoperimetric inequalities, cyclic inequalities, and a monotonic inequality for $L_p$-curvature images.

Introduction

In the study of convex geometry, researchers often seek transformations—or "images"—that reveal the fundamental properties of a convex body. A landmark contribution to this field was Erwin Lutwak's introduction of the $L_p$-curvature image. This transformation captures curvature information in a manner compatible with $L_p$ spaces, which are essential mathematical tools for measuring distance and size.

While Lutwak established a basic volume inequality for these images, a more comprehensive framework was required to understand their deeper geometric behavior, particularly their interaction with affine transformations. This paper, published in the Journal of Mathematical Inequalities, addresses this need by linking $L_p$-curvature images to $L_p$-affine and dual affine surface areas.

Methodology and Theoretical Context

The authors, Bin Chen and Weidong Wang, utilize the $L_p$-Brunn-Minkowski-Firey theory, a modern extension of classical geometry. Their approach involves applying sophisticated metrics—specifically $L_p$-affine surface area and $L_p$-dual affine surface area—to the curvature images. These metrics are particularly powerful because they are "affine-invariant," meaning they remain unchanged even if the shape is stretched or sheared.

By bridging the study of curvature images with affine surface area measures, the authors were able to construct formal analytical proofs for three distinct categories of mathematical inequalities: isoperimetric, cyclic, and monotonic.

Key Results and Insights

Affine Isoperimetric Inequalities

The researchers established affine isoperimetric inequalities for $L_p$-curvature images. These provide a lower bound for the "surface area" of the image relative to its volume. This extends the classical isoperimetric problem—which identifies the shape with the minimum perimeter for a given area—into the more complex realm of $L_p$ affine geometry.

Cyclic Inequalities

The paper proves that certain geometric quantities of these images satisfy a chain of cyclic inequalities. This suggests a recursive relationship between different "orders" of the images, implying a hidden hierarchy in how curvature information is structured within a convex body.

Monotonic Inequality

The authors demonstrated a monotonic inequality, showing that the $L_p$-curvature image behaves predictably as the parameter $p$ changes. This "one-way" evolution is crucial for establishing the stability of these geometric transformations across different $L_p$ metrics.

Conclusion and Future Directions

This work provides a robust theoretical expansion of $L_p$ geometry. By connecting curvature images to affine geometry, the authors have created new tools for bounding the properties of these complex transformations.

While the paper remains in the realm of pure mathematics, its findings have potential implications for fields such as computer vision and materials science, where convex geometry is frequently applied. Future research may extend these results to Orlicz spaces, which offer an even more general framework than $L_p$ spaces, or apply these inequalities to solve specific instances of the $L_p$-Minkowski problem, which seeks to reconstruct convex bodies from curvature data.

Summary updated

Abstract

Lutwak introduced the notion of L p -curvature image and proved an inequality for volumes of convex body and its L p -curvature image. In this article, based on the L p -affine surface area and L p -dual affine surface area, we establish the affine isoperimetric inequalities, cyclic inequalities and a monotonic inequality for L p -curvature images.
BibTeX
@article{chen2019some,
  author = {Bin Chen and Weidong Wang},
  title = {{Some inequalities for the L\_p-curvature images}},
  journal = {Journal of Mathematical Inequalities},
  year = {2019},
  doi = {10.7153/jmi-2019-13-05},
}
Figure