1: \begin{abstract} We develop mask iterative hard thresholding

2:   algorithms (mask IHT and mask DORE) for sparse image reconstruction

3:   of objects with known contour.  The measurements follow a noisy

4:   underdetermined linear model common in the compressive sampling

5:   literature.  Assuming that the contour of the object that we wish to

6:   reconstruct is \emph{known} and that the signal outside the contour

7:   is zero, we formulate a constrained residual squared error

8:   minimization problem that incorporates \emph{both} the geometric

9:   information (i.e.\ the knowledge of the object's contour) and the

10:   signal sparsity constraint. We first introduce a mask IHT method

11:   that aims at solving this minimization problem and guarantees

12:   monotonically non-increasing residual squared error for a given

13:   signal sparsity level. We then propose a double overrelaxation

14:   scheme for accelerating the convergence of the mask IHT algorithm.

15:   We also apply convex mask reconstruction approaches that employ a

16:   convex relaxation of the signal sparsity constraint.  In X-ray

17:   computed tomography (CT), we propose an \emph{automatic} scheme for

18:   extracting the convex hull of the inspected object from the measured

19:   sinograms; the obtained convex hull is used to capture the object

20:   contour information.  We compare the proposed mask reconstruction

21:   schemes with the existing large-scale sparse signal reconstruction

22:   methods via numerical simulations and demonstrate that, by

23:   exploiting both the geometric contour information of the underlying

24:   image and sparsity of its wavelet coefficients, we can reconstruct

25:   this image using a significantly smaller number of measurements than

26:   the existing methods.

27: \end{abstract}

28: