BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Sparse models and convex optimisation for convolutive blind source
  separation
DTSTART:20110308T140000
DTSTAMP:20260407T163541Z
UID:17d7eec4f9730e9f839f3230dce21810a33ca003a478c67b81d457ea
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dr. Prasad Sudhakar\, INRIA Rennes-Bretagne Atlantique\, Franc
 e\nWe jointly exploit the sparsity of the sources and mixing filters for b
 lind estimation of sparse filters from stereo convolutive mixtures of seve
 ral sources. First\, we show why the sparsity of the filters can help solv
 e the permutation problem in convolutive source separation\, in the absenc
 e of scaling. Then\, we propose a  two-stage estimation framework\, which 
 is primarily based on the time-frequency domain cross-relation and an $ell
 ^1$ minimisation formulation: a) a clustering step to group the time-frequ
 ency points where only one source is active\, for each source\; b) a conve
 x optimisation step which estimates the filters. The resulting algorithms 
 are assessed on audio source separation and filter estimation problems.
LOCATION:CO 016 https://plan.epfl.ch/?room==CO%20016
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR