BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:Australia/Perth
BEGIN:STANDARD
DTSTART:20090329T020000
TZOFFSETFROM:+0900
TZOFFSETTO:+0800
TZNAME:AWST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20230323T025758Z
UID:9F39B76E-5722-4B17-8CFF-CC12730C8C07
DTSTART;TZID=Australia/Perth:20230307T083000
DTEND;TZID=Australia/Perth:20230307T100000
DESCRIPTION:Blind source separation (BSS) separates mixtures into individua
 l sources with as little prior information as possible. Independent compon
 ent analysis (ICA) is the primary method for BSS that estimates a linear t
 ransformation for separating a mixture into statistically independent comp
 onents. ICA has many applications\, including image and biomedical signal 
 processing. For audio sources such as speech and music mixed in a real rev
 erberant environment\, we need to solve multiple ICA problems\, each corre
 sponding to a frequency of the audio signals. In addition\, we need to rou
 ghly estimate individual sources to align the permutation indeterminacy of
  the ICA solutions. Nonnegative matrix factorization (NMF) provides a path
 way to model each sound source by a low-rank approximation. NMF identifies
  frequent sound patterns in spectrograms. In this presentation\, Dr Hirosh
 i Sawada (NTT communication Science Laboratories) will address combined NM
 F and ICA for simultaneous estimating of individual sources and linear tra
 nsformations. Their reserach developed a sophisticated independent low-ran
 k matrix analysis (ILRMA) method for audio BSS tasks. A live demonstration
  of BSS to separate two simultaneous speech sources recorded with a stereo
  IC recorder\, will be provided.\n\nSpeaker(s): Dr Hiroshi Sawada\, \n\nBe
 ntley\, Western Australia\, Australia\, Virtual: https://events.vtools.iee
 e.org/m/349573
LOCATION:Bentley\, Western Australia\, Australia\, Virtual: https://events.
 vtools.ieee.org/m/349573
ORGANIZER:syed.islam@ecu.edu.au
SEQUENCE:4
SUMMARY:IEEE SP Distinguished lecture\, "Simultaneous Estimation of Individ
 ual Sources and Linear Transformations: A Case of Blind Audio Source Separ
 ation" 
URL;VALUE=URI:https://events.vtools.ieee.org/m/349573
X-ALT-DESC:Description: <br /><p>Blind source separation (BSS) separates mi
 xtures into individual sources with as little prior information as possibl
 e. Independent component analysis (ICA) is the primary method for BSS that
  estimates a <em>linear transformation</em> for separating a mixture into 
 statistically independent components. ICA has many applications\, includin
 g image and biomedical signal processing. For audio sources such as speech
  and music mixed in a real reverberant environment\, we need to solve mult
 iple ICA problems\, each corresponding to a frequency of the audio signals
 . In addition\, we need to roughly estimate <em>individual sources</em> to
  align the permutation indeterminacy of the ICA solutions. Nonnegative mat
 rix factorization (NMF) provides a pathway to model each sound source by a
  low-rank approximation. NMF identifies frequent sound patterns in spectro
 grams. In this presentation\, Dr Hiroshi Sawada (NTT communication Science
  Laboratories) will address combined NMF and ICA for simultaneous estimati
 ng of individual sources and linear transformations. Their reserach develo
 ped a sophisticated independent low-rank matrix analysis (ILRMA) method fo
 r audio BSS tasks. A live demonstration of BSS to separate two simultaneou
 s speech sources recorded with a stereo IC recorder\, will be provided.</p
 >
END:VEVENT
END:VCALENDAR