Uniformly Perfect Measures

• Uniformly perfect measures are defined as Borel measures on metric spaces that satisfy a reverse-doubling condition, ensuring a minimum mass gain when enlarging balls.
• The lower regularity dimension quantitatively captures the minimal power-law growth and characterizes uniform perfectness for doubling measures.
• These measures maintain stability under quasisymmetric mappings and play a key role in applications such as geometric analysis and Diophantine approximation.

A uniformly perfect measure is a locally finite Borel measure on a metric space that satisfies a quantitative reverse-doubling condition: enlarging a ball by a fixed factor must increase its mass by at least a fixed multiplicative amount. In the regularity-dimension framework, this notion is encoded by the lower regularity dimension, which measures the sharp lower power-law growth of ball masses across scales. For doubling measures, uniform perfectness is equivalent to positivity of the lower regularity dimension, so the subject sits at the intersection of quantitative doubling theory, Assouad-type dimensions, quasisymmetric geometry, and applications such as Diophantine approximation (Howroyd, 2019).

1. Definitions and foundational framework

Let $(X,d)$ be a metric space and $\mu$ a locally finite Borel measure on $X$. The upper comparison notion is doubling: $\mu$ is doubling if there exists $C(2)>1$ such that

$\mu(B(x,R)) \le C(2)\,\mu(B(x,R/2))$

for all $x\in X$ and all $R>0$. More generally, if $X$ is a doubling space, for each fixed $\theta>1$ one may write

$\mu$0

The measure-theoretic analogue of uniform perfectness reverses this inequality in scale. A measure $\mu$1 is uniformly perfect, or reverse-doubling, if there exists $\mu$2 such that

$\mu$3

for all $\mu$4 and all $\mu$5. Equivalently, for every $\mu$6 there exists $\mu$7 with

$\mu$8

This requires that passing from radius $\mu$9 to radius $X$0 produces a definite gain in mass, uniformly in the center and scale (Howroyd, 2019).

The relevant quantitative invariant is the lower regularity dimension,

$X$1

Its upper counterpart is the upper regularity dimension,

$X$2

These quantities are measure-theoretic Assouad-type dimensions. When both inequalities are available, they package local growth into the two-sided estimate

$X$3

2. Characterization by lower regularity dimension

The basic structural fact is that, for doubling measures, uniform perfectness is exactly positivity of the lower regularity dimension. More precisely, the equivalence established in the regularity-dimension framework states that for a doubling measure $X$4,

$X$5

Thus uniform perfectness is not merely a qualitative anti-degeneracy condition; it is the statement that mass growth has a strictly positive lower power exponent across all locations and scales (Howroyd, 2019).

The lower regularity dimension also recovers the optimal reverse-doubling exponent. If $X$6 is uniformly perfect with reverse-doubling constants $X$7, then

$X$8

This is the exact analogue of the formula expressing the upper regularity dimension in terms of doubling constants. It shows that $X$9 is the best exponent compatible with all lower scaling inequalities.

A further quantitative result links measure-theoretic uniform perfectness to geometric uniform perfectness of the ambient space. If $\mu$0 is uniformly perfect and $\mu$1 is a doubling, fully supported measure on $\mu$2, then $\mu$3 is uniformly perfect. In fact, if $\mu$4 is $\mu$5-uniformly perfect and $\mu$6 has doubling constants $\mu$7, then

$\mu$8

This gives an explicit lower bound for the lower regularity dimension in terms of the space’s uniform perfectness constant and the measure’s doubling data (Howroyd, 2019).

At the same time, there is no universal functional relation between $\mu$9 and $C(2)>1$0 alone. The upper regularity dimension depends only on doubling constants, whereas the lower regularity dimension is constrained by the lower dimension of the underlying space. Consequently, the geometry of the support enters essentially into the lower theory.

3. Ambient geometry and existence of uniformly perfect measures

Uniform perfectness of the underlying space has strong consequences for all doubling measures it carries. In a bounded uniformly perfect metric space, every doubling measure satisfies two-sided power estimates on balls: there exist $C(2)>1$1 such that

$C(2)>1$2

The lower bound comes from iterating doubling downward in scale, while the upper bound uses uniform perfectness together with doubling in a reverse-doubling form. This places every doubling measure on a bounded uniformly perfect space within a power-growth regime (Ojala et al., 2011).

The existence theory for lower-regular measures is correspondingly broad. If $C(2)>1$3 is a uniformly perfect complete metric space satisfying the finite doubling property, then for every $C(2)>1$4 there exists a fully supported measure $C(2)>1$5 such that

$C(2)>1$6

where $C(2)>1$7 is the lower dimension of $C(2)>1$8. If $C(2)>1$9 is in addition doubling, $\mu(B(x,R)) \le C(2)\,\mu(B(x,R/2))$0 can be chosen doubling as well. The theorem is constructive, using a nested cube structure and a scale-by-scale assignment of mass that forces the lower growth estimate defining $\mu(B(x,R)) \le C(2)\,\mu(B(x,R/2))$1 (Käenmäki et al., 2016).

A useful corollary identifies the metric and measure-theoretic notions sharply: for complete metric spaces with the finite doubling property,

$\mu(B(x,R)) \le C(2)\,\mu(B(x,R/2))$2

This equivalence makes lower regularity dimension the natural measure-theoretic realization of the lower dimension of the space. The existence theorem gives approximation from below, but it does not in general assert exact equality $\mu(B(x,R)) \le C(2)\,\mu(B(x,R/2))$3.

4. Stability under mappings and failure under irregular parametrizations

Uniformly perfect measures behave well under quasisymmetric homeomorphisms. If $\mu(B(x,R)) \le C(2)\,\mu(B(x,R/2))$4 is uniformly perfect, $\mu(B(x,R)) \le C(2)\,\mu(B(x,R/2))$5 is doubling on $\mu(B(x,R)) \le C(2)\,\mu(B(x,R/2))$6, and $\mu(B(x,R)) \le C(2)\,\mu(B(x,R/2))$7 is an $\mu(B(x,R)) \le C(2)\,\mu(B(x,R/2))$8-quasisymmetric homeomorphism with

$\mu(B(x,R)) \le C(2)\,\mu(B(x,R/2))$9

then for the pushforward measure $x\in X$0,

$x\in X$1

and similarly

$x\in X$2

Thus quasisymmetric maps preserve both doubling and positivity of lower regularity dimension, with quantitative control. In particular, uniformly perfect measures are quasisymmetric invariants in the measure-theoretic sense (Howroyd, 2019).

The opposite behavior occurs for pushforwards onto graphs of random processes. Let $x\in X$3 be a doubling measure on $x\in X$4, let $x\in X$5 be an $x\in X$6-stable Lévy process with non-vanishing distribution for $x\in X$7, and let $x\in X$8. Then almost surely the pushforward $x\in X$9 on the graph $R>0$0 is neither doubling nor uniformly perfect; equivalently,

$R>0$1

This shows that regularity dimensions are highly sensitive to the map, not merely to the image set. The graph itself can have large Assouad dimension while carrying natural pushforward measures with vanishing lower regularity dimension.

The contrast isolates the structural role of quasisymmetry. Controlled distortion of relative distances preserves reverse-doubling exponents, whereas random graph parametrizations can destroy them completely.

5. Decay conditions and Diophantine applications

Uniformly perfect measures are closely tied to weak absolute decay. A measure $R>0$2 is weakly absolutely $R>0$3-decaying if there exist $R>0$4 such that for all $R>0$5, all $R>0$6, and all $R>0$7,

$R>0$8

This is an upper estimate for shrinking the radius, but after rearrangement it becomes a lower growth estimate for enlarging the radius.

The precise link is: $R>0$9 provided $X$0. Hence, for doubling measures, uniform perfectness is equivalent to weak absolute $X$1-decay for all $X$2. The lower regularity dimension is therefore the optimal decay exponent in this class (Howroyd, 2019).

This perspective feeds directly into Diophantine approximation for Kleinian groups. In the Beresnevich–Ghosh–Simmons–Velani theorem, if $X$3 is a nonelementary geometrically finite Kleinian group, $X$4 is a parabolic or hyperbolic fixed point, $X$5 is compact, and $X$6 is a weakly absolutely $X$7-decaying measure on $X$8, then

$X$9

Using the characterization above, the theorem can be reformulated in terms of lower dimension of the support and lower regularity dimension of the measure. In particular, Patterson–Sullivan measures on the limit set have strictly positive and finite regularity dimensions, hence are uniformly perfect and weakly absolutely $\theta>1$0-decaying for all $\theta>1$1. This makes lower regularity dimension the relevant exponent controlling admissible decay rates in Diophantine estimates.

6. Broader context and adjacent notions

Uniformly perfect measures are the measure-theoretic counterpart of uniformly perfect sets. On the set side, a metric space is uniformly perfect if and only if it has positive lower dimension, and for uniformly perfect sets in doubling metric spaces the lower Assouad-type dimensions remain strictly positive and admit a refined spectrum theory (Chen et al., 2018). In planar hyperbolic geometry, uniform perfectness of $\theta>1$2 is equivalent to positivity of a hyperbolic-density–distance product, namely

$\theta>1$3

with an equivalent formulation using $\theta>1$4 (Sugawa, 2015). These set-theoretic criteria clarify the geometric background against which reverse-doubling measures arise.

Uniformly perfect measures should also be distinguished from stronger uniformity classes. An $\theta>1$5-uniform measure on $\theta>1$6 satisfies

$\theta>1$7

for all $\theta>1$8 and all $\theta>1$9; such measures are much more rigid than merely reverse-doubling, and in Tolsa’s theory they have big pieces of Lipschitz graphs and are uniformly rectifiable (Tolsa, 2013). In the one-dimensional metric setting, 1-uniform spaces satisfy $\mu$00 exactly and there are only three models: $\mu$01, $\mu$02, and $\mu$03 (Bate, 10 Jan 2025). These are adjacent extremal notions rather than variants of uniform perfectness.

The support geometry alone does not determine uniform perfectness. There exist compact hereditarily non uniformly perfect sets in the plane of Hausdorff dimension $\mu$04 and positive logarithmic capacity, and also HNUP sets of positive Lebesgue measure (Stankewitz et al., 2016). This indicates that neither Hausdorff dimension, nor capacity, nor measure-theoretic largeness of the support suffices to guarantee reverse-doubling behavior. Uniformly perfect measures therefore occupy a genuinely intermediate position: weaker than exact regularity or Ahlfors regularity, but stronger than mere largeness or positive dimension of the support.

In this framework, the lower regularity dimension is the central invariant. It quantifies how strongly a measure avoids small-scale concentration, identifies uniform perfectness for doubling measures, behaves naturally under quasisymmetric maps, interacts with weak absolute decay, and links the measure theory of a support to its lower-dimensional geometry.