Inequality of Hausdorff measures for convex sets $\mathfrak{H}^{n-1}(\partial E)\le \mathfrak{H}^{n-1}(\partial F)$

geometric-measure-theorygeometryhausdorff-measurereal-analysis

I'm preparing for an exam on the calculus of variations and I need help in solving this exercise from an old exam text (actually it's only a part of a bigger exercise but its parts are quite independent): I don't know where to start and any help would be welcomed.

Given two nonempty convex subsets $E\subset F\subset \mathbb{R}^n$, $n\ge 2$ , prove that
$\mathfrak{H}^{n-1}(\partial E)\le \mathfrak{H}^{n-1}(\partial F).$
(In fact, the convexity of $F$ is not needed.)

Here $\partial E$ is the boundary of the set $E$ and $\mathfrak{H}$ is the standard Hausdorff measure i.e. the measure defined as
$$\newcommand{\diam}{\operatorname{diam}}
\mathfrak{H}^d(E)=\liminf_{\underset{\{U_i\}_{i}:\bigcup_{i}U_i\supseteq E}{{\delta\to 0}}}\bigg\{ \sum_{i=1}^\infty (\diam U_i)^d:\bigcup_{i=1}^\infty U_i\supseteq E,\: \diam(U_i)<\delta \bigg\} $$
where we are working in $\mathbb{R}^n$ seen as a metric space $(X,p)$ and $$ \diam(U_i)=\sup\{p(x,y):x,y\in U_i\}$$ as it's defined in any undergrad analysis course dealing with multi-variate integration.

My problem is that I never studied geometric measure theory, except for the basic knowledge I acquired in the introductory course in multivariate integration, so I'm finding it hard to find a way from the hypothesis of convexity of the involved sets through its application to the definition of Hausdorff measure.

Best Answer

It's enough to do a projection from $\partial(F)$ to $\partial E$.

This is a Lipshitz function with Lipshitz constant 1 and so by the properties of the Hausdorff measure we know that $H^s(f(E))\le L^s (H^s(E))$ and so we are done