BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Fast L0-Based Sparse Signal Recovery
DTSTART:20110117T133000
DTSTAMP:20260413T094550Z
UID:2ee31a93e120e1ab0f97ecc69f1130043dc8a0d883ac89af37a056ec
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Nick Kingsbury\, Signal Processing and Communications Re
 search Group\, University of Cambridge\nWe develop an algorithm for recons
 tructing signals from limited observations in a linear system. We assume a
 n adaptive Gaussian model for the signals\, which are assumed to be sparse
  in some suitable domain (e.g. the complex wavelet domain). This model res
 ults in a least squares problem with an iteratively reweighted L2 penalty 
 that is shown to approximate the L0-norm. We propose a fast algorithm to s
 olve the problem within a continuation framework\, in order to encourage c
 onvergence to the globally optimum solution. In our examples\, we show tha
 t the correct sparsity map and sparsity level are gradually learnt during 
 the iterations even when the number of observations is reduced\, or when o
 bservation noise is present. In addition\, with the help of interscale sig
 nal models\, the algorithm is able to recover signals to a better accuracy
  and with reduced numbers of observations\, when compared with L1 and rewe
 ighted-L1 methods.  The algorithm is applicable to a wide range of input d
 ata\, including high-resolution 2D images or 3D datasets and can be used t
 o achieve various forms of super-resolution.
LOCATION:CM 1 3 https://plan.epfl.ch/?room==CM%201%203
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR