Performance of a nullspace-map image reconstruction algorithm.
OPTICAL TOMOGRAPHY AND SPECTROSCOPY OF TISSUE: THEORY, INSTRUMENTATION, MODEL, AND HUMAN STUDIES II, PROCEEDINGS OF
185 - 196.
There are two reasons that might be attributed to the difficulty for the imaging problem in optical tomography, and in inverse problems in general. Firstly: the problem is mostly underdetermined. Secondly, the inverse problem is highly ill-conditioned due to the diffusive nature of the photons. We introduce Bayesian optimization that provides a method to incorporate a priori knowledge in the inversion and we show with the concept of nullspace that the Bayesian prior probability generalizes conventional regularization by introducing a prior model. Reconstruction results of test objects from simulated data and a reconstruction example on a head model show that use the nullspace gives considerable improvement.
|Title:||Performance of a nullspace-map image reconstruction algorithm|
|Location:||SAN JOSE, CA|
|Keywords:||optical tomography, image reconstruction, maximum a posteriori probability, ill-posed problems, regularization, Bayesian probability, prior knowledge|
|UCL classification:||UCL > School of BEAMS > Faculty of Engineering Science
UCL > School of BEAMS > Faculty of Engineering Science > Computer Science
UCL > School of BEAMS > Faculty of Engineering Science > Medical Physics and Bioengineering
Archive Staff Only