BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//University of Liverpool Computer Science Seminar System//v2//EN
BEGIN:VEVENT
DTSTAMP:20260610T182948Z
UID:Seminar-ACTO-980@lxserverM.csc.liv.ac.uk
ORGANIZER:CN=Nikhil Mande:MAILTO:Nikhil.Mande@liverpool
DTSTART:20210127T140000
DTEND:20210127T150000
SUMMARY:Algorithms, Complexity Theory and Optimisation Series
DESCRIPTION:Andreas Alpers: On the computational complexity of super-resolution imaging in discrete tomography\n\nSuper-resolution imaging aims at improving the resolution of\n\nan image by enhancing it with other images or data that might have been\n\nacquired using different imaging techniques or modalities. Motivated by\n\napplications in plasma physics, we consider the task of doubling the\n\nresolution of tomographic grayscale images of binary objects by fusion\n\nwith double-resolution tomographic data that has been acquired from two\n\nviewing angles. We show that this task is polynomial-time solvable if\n\nthe gray levels have been reliably determined. The task becomes NP-hard\n\nif the gray levels of some pixels come with an error of +-1 or larger.\n\nThe NP-hardness persists for any larger resolution enhancement factor.\n\nThis means that noise does not only affect the quality of a\n\nreconstructed image but, less expectedly, also the algorithmic\n\ntractability of the inverse problem itself. (This is joint work with\n\nPeter Gritzmann, TU Munich.)\n\nhttps://www.csc.liv.ac.uk/research/seminars/abstract.php?id=980
LOCATION:Online MT CSACTOO365Team
END:VEVENT
END:VCALENDAR