Robust and stable region-of-interest tomographic reconstruction by sparsity-inducing convex optimizationMS37

We propose an improved framework for region-of-interest computed tomography in the situation of noisy projection data. Under the assumption of a robust width prior that generalizes sparsity norms and measurement models used in compressed sensing, we derive performance guarantees by establishing error bounds for robust and stable recovery. Our reconstruction algorithm is guaranteed to satisfy predetermined fidelity and consistency tolerances while controlling the reconstruction error. It performs very competitively with respect to state-of-the-art methods.

This presentation is part of Minisymposium “MS37 - Sparse-based techniques in variational image processing (2 parts)
organized by: Serena Morigi (Dept. Mathematics, University of Bologna) , Ivan Selesnick (New York University) , Alessandro Lanza (Dept. Mathematics, University of Bologna) .

Demetrio Labate (University of Houston)
computed tomography, inverse problems, nonlinear optimization