BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:Poland
BEGIN:DAYLIGHT
DTSTART:20130331T030000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20131027T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20140131T204157Z
UID:F0254977-E5B6-11E7-833E-0050568D7F66
DTSTART;TZID=Poland:20130925T120000
DTEND;TZID=Poland:20130925T140000
DESCRIPTION:A problem of processing of I-vectors in the text-independent sp
 eaker verification systems was discussed. A new generalized cosine similar
 ity optimization technique was proposed. The optimization was performed ov
 er sets of orthogonal and diagonal matrices. Lie group techniques were use
 d in order to obtain exact orthogonal matrix solution. The experiments hav
 e been performed using the NIST2010 database. It was shown that a substant
 ial equal error rate (EER) reduction can be obtained with the proposed alg
 orithm.\n\nPoznan\, Wielkopolskie\, Poland
LOCATION:Poznan\, Wielkopolskie\, Poland
ORGANIZER:
SEQUENCE:0
SUMMARY:[Legacy Report] Generalized cosine similarity in I-vector based spe
 aker recognition systems
URL;VALUE=URI:https://events.vtools.ieee.org/m/98469
X-ALT-DESC:Description: <br />A problem of processing of I-vectors in the t
 ext-independent speaker verification systems was discussed. A new generali
 zed cosine similarity optimization technique was proposed. The optimizatio
 n was performed over sets of orthogonal and diagonal matrices. Lie group t
 echniques were used in order to obtain exact orthogonal matrix solution. T
 he experiments have been performed using the NIST2010 database. It was sho
 wn that a substantial equal error rate (EER) reduction can be obtained wit
 h the proposed algorithm.
END:VEVENT
END:VCALENDAR