Variational Convergence Analysis With Smoothed-TV Interpretation

Publication date 16 November 2015

Abstract: The problem of minimizing the least squares functional with a Fr'echet differentiable, lower semi-continuous, convex penalizer

J

is considered to be solved. The penalizer maps the functions of Banach space

m a t h c a l V

into

m a t h b b R_{+},

J : m a t h c a l V i g h t a r r o w m a t h b b R_{+} .

It is assumed that some given data

f^{d e l t a}

is defined on a compact domain

m a t h c a l G s u b s e t m a t h b b R_{+}

and in the class of Hilbert space,

f^{d e l t a} i n m a t h c a l L^{2} (m a t h c a l G) .

Then general Tikhonov functional associated with some given linear, compact and injective forward operator

m a t h c a l T : m a t h c a l V i g h t a r r o w m a t h c a l L^{2} (m a t h c a l G)

is formulated as �egin{eqnarray} F_{alpha}(varphi, f^{delta}) : & mathcal{V} imes mathcal{L}^{2}(mathcal{G}) & ightarrow mathbb{R}_{+}

onumber\ & (varphi, f^{delta}) & mapsto F_{alpha}(varphi, f^{delta}) := frac{1}{2}Vertmathcal{T}varphi - f^{delta}Vert_{mathcal{L}^{2}(mathcal{G})}^2 + alpha J(varphi) .

onumber end{eqnarray} Convergence of the regularized solution

v a r p h i_{a l p h a (d e l t a)} i n m a t h r m {a r g m i n}_{v a r p h i i n m a t h c a l V} F_{a l p h a} (v a r p h i, f^{d e l t a})

to the true solution

v a r p h i^{d a g g e r}

is analysed by means of Bregman divergence. First part of this aims to provide some general convergence analysis for generally strongly convex functional

J

in the cost functional

F_{a l p h a}

. In this part the key observation is that strong convexity of the penalty term

J

with its convexity modulus implies norm convergence in the Bregman metric sense. In the second part, this general analysis will be interepreted for the smoothed-TV functional. The result of this work is applicable for any strongly convex functional.

This page was built for publication: Variational Convergence Analysis With Smoothed-TV Interpretation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6267472)