Sparse X-ray tomography using Bayesian inversion

Date(s) : 19/02/2014   iCal
14 h 00 min - 15 h 00 min

$“A\; sparsity\; promoting\; reconstruction\; method\; is\; studied\; in\; the\; context\; of\; X-ray\; tomography\; with\; limited\; X-ray\; projection\; data.\; The\; reconstruction\; method\; is\; based\; on\; minimizing\; a\; sum\; of\; \$\; l^2\$-norm\; and\; a\; \$l^1\$-norm.\; Especially\; considered\; is\; the\; \$l^1\$-norm\; of\; wavelet\; coefficients.\; Depending\; on\; the\; viewpoint\; this\; method\; can\; be\; considered\; either\; a)\; as\; the\; inverse\; problem\; of\; finding\; a\; Bayesian\; MAP\; estimate\; with\; Besov\; \$B\_\{11\}^1\$\; space\; prior\; or\; b)\; as\; a\; deterministic\; regularization\; with\; Besov\; norm\; penalty.$
A tailored large-scale primal-dual interior-point method is used to solve the associated constrained minimization problem. The selection of the regularization parameter (or prior parameter, depending on the viewpoint) is performed by a novel technique called the S-curve method. Numerical results are presented both from simulated and from real, experimental data.”

Catégories

• GdT Analyse Spectrale et Problèmes Inverses (ASPI)