On the packing dimension of projected measures
Nicolas Angelini · 2026
Read On the packing dimension of projected measures by Nicolas Angelini (2026) 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-2604-18222.txt.
Introduction
Angelini introduces a threshold for Borel probability measures that governs when projections keep full packing dimension. For each level up to the ambient one, a typical projection keeps it exactly when the level does not exceed the threshold. Beyond that point, the Assouad dimension of the support controls how much can drop. When the level reaches the Assouad dimension, the full value is kept for nearly every subspace, recovering a known set result in Assouad form.
Catalog record
| CREATOR | Nicolas Angelini |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 30181 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.18222
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
On the packing dimension of projected measures Nicolas Angelini Abstract . We study the packing dimension of Borel measures under orthogonal projections. We give a necessary and sufficient condition such that typical projections of Borel probability measures have full packing dimension and derive general lower bounds in the complementary case. Our approach shows that the Assouad dimension of the support influences the behavior of projected measures. The same method yields corresponding results for images under fractional Brownian motion. 2010 Mathematics Subject Classification Primary 2010 Mathematics Subject Classification Primary 28A78, 28A75 1. Introduction Let \mu be a finite Borel measure on \mathbb{R}^{n} , and for V\in G(n,m) let \pi_{V} denote the orthogonal projection onto V . We write \mu_{V}=\mu\circ\pi_{V}^{-1} for the projected measure. For Hausdorff dimension, it is classical that (1.1) \dim_{H}\mu_{V}=\min\{\dim_{H}\mu,m\} for \gamma_{n,m} -almost every V\in G(n,m) , see [ 6 ] . In contrast, the behavior of packing dimension under projections is more subtle. Falconer and Howroyd introduced packing dimension profiles and proved that for any finite Borel measure \mu on \mathbb{R}^{n} , \dim_{P}\mu_{V}\leq\dim_{P}^{m}\mu for all V\in G(n,m) , with equality for \gamma_{n,m} -almost every V , [ 5 ] . This shows that the packing dimension of typical projections is almost surely constant and determined by the m -dimensional profile of \mu . However, this result does not directly yield analogues of ( 1.1 ). It is straightforward to see that \dim_{P}\mu_{V}\leq\min\{\dim_{P}\mu,m\} for all V\in G(n,m) . On the other hand, Falconer and Mattila obtained a general lower bound: if \mu is a Borel probability measure on \mathbb{R}^{n} , then for \gamma_{n,m} -almost all V , \dim_{P}\mu_{V}\geq\begin{cases}\dfrac{\dim_{P}\mu\left(1-\frac{1}{n}\dim_{H}\mu\right)}{1+\left(\frac{1}{m}-\frac{1}{n}\right)\dim_{P}\mu-\frac{1}{m}\dim_{H}\mu},&\text{if }\dim_{H}\mu\leq m,\\[11.99998pt] m,&\text{if }\dim_{H}\mu\geq m,\end{cases} and they showed that this bound is sharp in general [ 4 ] . The main purpose of this article is to identify conditions under which Hausdorff-type projection formulas can be recovered for packing dimension. We introduce a parameter \mathcal{D}(\mu) (Definition 4.1 ) that governs when projections have full dimension, in the sense that \dim_{P}\mu_{V}=m\quad\text{for }\gamma_{n,m}\text{-a.e. }V\quad\Longleftrightarrow\quad m\leq\mathcal{D}(\mu), see Theorem 4.5 . This provides a complete characterization of the regime in which typical projections attain maximal dimension. To obtain lower bounds in the complementary case, \mathcal{D}(\mu)<m , we relate projection dimensions to the Assouad dimension of the support of \mu . Our main estimate (Theorem 4.12 ) asserts that \dim_{P}\mu_{V}\geq\dim_{P}\mu-\max\{0,\dim_{A}(\supp\mu)-m\} for \gamma_{n,m} -almost every V\in G(n,m) . In particular, if m\geq\dim_{A}(\supp\mu) , then \dim_{P}\mu_{V}=\dim_{P}\mu for typical projections. This shows that the Assouad dimension of the support controls the stability of packing dimension under orthogonal projections. Related results for packing and box dimensions of sets were obtained in [ 3 ] , and analogous results for the \theta -intermediate dimension were obtained in [ 1 ] , see also [ 2 ] . In [ 9 ] , Xiao extended the theory of packing dimension profiles to random images of measures under fractional Brownian motion. He proved that if X:\mathbb{R}^{n}\to\mathbb{R}^{d} is a fractional Brownian motion of index \alpha , see [ 7 , Chapter 18] , \mu is a finite Borel measure on \mathbb{R}^{n} , and \mu_{X} denotes the measure defined by \mu_{X}(B)=\mu\{y\in\mathbb{R}^{n}:X(y)\in B\} for any Borel set B\subset\mathbb{R}^{d} , then \dim_{P}\mu_{X}=\frac{1}{\alpha}\dim_{P}^{\alpha d}\mu, with probability 1 . Since our results for orthogonal projections are obtained using packing dimension profiles, our estimates also apply to the setting of fractional Brownian motion. That is, with the same notation as above, if \alpha d\geq\dim_{A}(\supp\mu) , then \dim_{P}\mu_{X}=\frac{\dim_{P}\mu}{\alpha}, and \dim_{P}\mu_{X}=d\hfill\iff\hfill\alpha d\leq\mathcal{D}(\mu). 2. Preliminaries We write B(x,r) to the ball of center x and radious r and for a bounded set E\subset\mathbb{R}^{n} we write |E| for its diameter. Throughout the paper, \mu denotes a Borel probability measure on \mathbb{R}^{n} , i.e. \mu(\mathbb{R}^{n})=1 . We write \supp\mu for the support of \mu . For 1\leq m\leq n , we write G(n,m) for the Grassmannian of m -dimensional linear subspaces of \mathbb{R}^{n} , endowed with its natural invariant probability measure \gamma_{n,m} , see [ 8 , Chapter 3] . For V\in G(n,m) , let \pi_{V}:\mathbb{R}^{n}\to V denote the orthogonal projection onto V , and define the projected measure by \mu_{V}=\mu\circ\pi_{V}^{-1}. This is a finite Borel measure on V . Hausdorff dimension and s -dimensional Hausdorff measure will be denoted by \dim_{H} and \mathcal{H}^{s} , respectively. The Assouad dimension of a set E\subset\mathbb{R}^{n} , \dim_{A}E is defined by \displaystyle\dim_{A}E:=\inf\Bigl\{s\geq 0:\ \displaystyle\exists C>0\text{ such that for all }0<r<R,\ x\in E, \displaystyle N_{r}\bigl(E\cap B(x,R)\bigr)\leq C\left(\frac{R}{r}\right)^{s}\Bigr\}. where N_{r}(E) denotes the minimal number of balls of radius r needed to cover E . Let 0<s<\infty , a closed set E\subset\mathbb{R}^{n} is called s -regular if there exists Borel measure \mu on \mathbb{R}^{n} and a constant C\geq 1 such that \mu(\mathbb{R}^{n}\setminus E)=0 and C^{-1}r^{s}\leq\mu\bigl(E\cap B(x,r)\bigr)\leq Cr^{s} for all x\in E and all 0<r<|E| , r<\infty . The packing dimension of a Borel measure \mu is defined by \dim_{P}\mu=\inf\{\dim_{P}E:E\text{ is a Borel set with }\mu(E)>0\}, where \dim_{P}E denote the packing dimension of the set E . An equivalent characterization is as follows, see [ 6 , Lemma 4.1] \dim_{P}\mu=\sup\left\{t\geq 0\,:\,\liminf_{r\to 0}\frac{\mu(B(x,r))}{r^{t}}=0\,\text{ for }\mu-a.e.x\in\mathbb{R}^{n}\right\}. After that, in [ 5 ] , Falconer and Howroyd gives a characterization of packing dimension in terms of the potential defined by F_{n}^{\mu}(x,r)=\int_{\mathbb{R}^{n}}\min\{1,r^{n}|x-y|^{-n}\}d\mu(y). Lemma 2.1 . [ 5 , Corollary 3] For \mu a finite Borel measure on \mathbb{R}^{n} we have \dim_{P}\mu=\sup\left\{t\geq 0\,:\,\liminf_{r\to 0}\frac{F_{n}^{\mu}(x,r)}{r^{t}}\text{ for }\mu-a.e.x\in\mathbb{R}^{n}\right\}. In the same article, in order to obtain sharp projection theorems, the authors introduced packing dimension profiles. For each integer 1\leq m\leq n , the m -dimensional packing dimension profile of a measure \mu is defined as follows. Definition 2.2 . For a finite Borel measure \mu on \mathbb{R}^{n} and 0\leq m\leq n define F_{m}^{\mu}(x,r)=\int_{\mathbb{R}^{n}}\min\{1,r^{m}|x-y|^{-m}\}d\mu(y) and \dim_{P}^{m}\mu=\sup\left\{t\geq 0\,:\,\liminf_{r\to 0}\frac{F_{m}^{\mu}(x,r)}{r^{t}}=0\text{ for }\mu-a.e.x\in\mathbb{R}^{n}\right\}. The fundamental result of Falconer and Howroyd asserts that for a finite Borel measure \mu on \mathbb{R}^{n} , the packing dimension of the projection \mu_{V} onto a typical m -dimensional subspace V is given by the m -dimensional packing dimension profile of \mu : Theorem 2.3 . [ 5 , Theorem 6] Let \mu be a finite Borel measure on \mathbb{R}^{n} . Then, for all V\in G(n,m) we have \dim_{P}\mu_{V}\leq\dim_{P}^{m}\mu with an equality for \gamma_{n,m}-a.e.V\in G(n,m) . The best lower bound was obtained in [ 4 ] by Mattila and Falconer Theorem 2.4 . [ 4 , Theorem 3.3] Let \mu be a Borel probability measure on \mathbb{R}^{n} . Then for \gamma_{n,m} almost all V\in G(n,m) , \dim_{p}\mu_{V}\geq\begin{cases}\dfrac{\dim_{p}\mu\left(1-\frac{1}{n}\dim_{H}\mu\right)}{1+\left(\frac{1}{m}-\frac{1}{n}\right)\dim_{p}\mu-\frac{1}{m}\dim_{H}\mu},&\text{if }\dim_{H}\mu\leq m,\\[11.99998pt] m,&\text{if }\dim_{H}\mu\geq m.\end{cases} and the authors shows with an example that this is the best possible lower bound. Moreover, in [ 5 ] an estimate for the size of the set of exceptional projection directions was also obtained. Theorem 2.5 . [ 5 , Theorem 13] Let \mu be a finite Borel Measure on \mathbb{R}^{n} and 0\leq s\leq m . Then for \mathcal{H}^{s+m(n-m-1)}- almost all V\in G(n,m) we have \dim_{P}\mu_{V}\geq\dim_{P}^{s}\mu. Finally, in [ 9 ] , Xiao shows that the packing dimension of the image measure \mu_{X} of \mu under the fractional Brownian motion of index \alpha\in(0,1) , X(t) , defined by \mu_{X}(B):=\mu\{t\in\mathbb{R}^{n}:X(t)\in B\}\quad\text{ for any Borel set }B\in\mathbb{R}^{d}, is almost surely related to packing dimension profiles. Theorem 2.6 . [ 9 , Theorem 3.1] Let \mu be any finite Borel measure on \mathbb{R}^{n} . Then with probability 1 ,