1: \begin{abstract}

2: Several applications in medical imaging and non-destructive material testing lead to inverse elliptic coefficient problems,

3: where an unknown coefficient function in an elliptic PDE is to be determined from partial knowledge of its solutions.

4: This is usually a highly non-linear ill-posed inverse problem, for which unique reconstructability results, stability estimates and 

5: global convergence of numerical methods are very hard to achieve.

6: 

7: The aim of this note is to point out a new connection between inverse coefficient problems and semidefinite programming that may help addressing these challenges.

8: We show that an inverse elliptic Robin transmission problem with finitely many measurements can be equivalently rewritten as 

9: a uniquely solvable convex non-linear semidefinite optimization problem.

10: This allows to explicitly estimate the number of measurements that is required to achieve a desired resolution, to derive an error estimate for noisy data, 

11: and to overcome the problem of local minima that usually appears in optimization-based

12: approaches for inverse coefficient problems.

13: \keywords{Inverse Problem \and Finitely many measurements \and Monotonicity \and Convexity \and Loewner order}

14: % \PACS{PACS code1 \and PACS code2 \and more}

15: \subclass{35R30 \and 90C22}

16: \end{abstract}