BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Source Separation for Nonlinear Mixtures: How and Why?
DTSTART:20131003T140000
DTEND:20131003T150000
DTSTAMP:20260407T033434Z
UID:121d9d5b8d68a1511f67063d83e18366d4738f835896ebe1c4b84f95
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Christian Jutten\, University Joseph Fourier\, Grenoble\
 , France\nBio: Christian Jutten is Professor of Statistical Signal Process
 ing in the Engineering Department of University Joseph Fourier of Grenoble
 . From 1993 through to 2007\, he served as director or deputy director of 
 his laboratory. From 2007 to 2010\, he was both director of the Signal and
  Image Processing Department (more than 100 academic\, post-doc and PhD st
 udents) and deputy director of GIPSA-lab (about 300 academic\, post-doc an
 d PhD students\, and engineering and administrative staffs). His main rese
 arch contributions are in statistical signal processing\, and is well know
 n for his pioneer works in source separation since middle of 80's. He is a
 uthor or co-author of more than 75 papers in international journals\, 4 bo
 oks\, 25 keynote plenary talks and more than 150 communications in interna
 tional conferences. He is currently Deputy Director in charge of Signal an
 d Image Processing in National Institute of Information Sciences and its I
 nteractions in the French National Center for Scientific Research (CNRS). 
 He was member of IEEE Technical Committees : Blind Signal Processing of th
 e Circuits and Systems Society (2000 to now) and Machine Learning for Sign
 al Processing of the Signal Processing Society (2007 to 2012). He received
  the best paper award of Eurasip Signal Processing(1991) and IEEE Trans. o
 n Geoscience and Remote Sensing (2012)\, the Medal Blondel from the French
  Electrical Engineering Society (SEE) in 1997)\, and has been elected as a
  Senior Member of Institut Universitaire de France and IEEE Fellow in 2008
 \, and EURASIP fellow in 2013. In 2012\, he was a recipient of an ERC Adva
 nced Grant on challenges in extraction and separation of source (CHESS).\n
 The problem of source separation has been addressed mainly for linear mixt
 ures\, either memoryless or convolutive. Methods for solving the problem a
 re based on source assumptions like statistical independence (ICA)\, time 
 properties (coloration or nonstationarity)\, positivity or sparsity. Howev
 er\, although linearity is very often a convenient approximation\, there a
 re some applications in which the mixing process is clearly nonlinear.\nIn
  this talk\, in a first part\, I explain basics on source separation and m
 ain results in the linear case before pointing out the main problems encou
 ntered by source separation in nonlinear mixtures and how they can be over
 come.\nThen\, in a second part\, I will consider actual strongly nonlinear
  problems: one in image processing and another one in chemical sensor arra
 y processing. For each problem\, I will derive the nonlinear models\, show
  how source separation can be applied and experiment results which can be 
 achieved.
LOCATION:CM 1105 http://plan.epfl.ch/?lang=fr&room=CM+1105
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR