H93-1019 |
98.5 % . With 2.5 s of speech the
|
speaker identification
|
accuracy is 98.3 % . For the
|
H93-1019 |
using unsupervised adaptation for
|
speaker identification
|
on the 168 TIMIT speakers had
|
H93-1019 |
experimented with this approach for
|
speaker identification
|
. In our implementation , each
|
H93-1019 |
of sentences longer than 3s ,
|
speaker identification
|
was correct , suggesting that
|
H93-1019 |
Table 5 , for text-independent
|
speaker identification
|
. As for sex and language identification
|
H90-1103 |
Speaker Verification ( SV ) 2.2
|
Speaker Identification
|
( SI ) 2.3 Language Identification
|
H93-1013 |
speaker-dependent acoustic models for
|
speaker identification
|
. The authors report good performance
|
H93-1019 |
effective for language , sex , and
|
speaker identification
|
and can enable better and more
|
H93-1019 |
text-independent language , sex , and
|
speaker identification
|
and can enable better and more
|
H01-1051 |
nonlexical information such as
|
speaker identification
|
and characterization , voice
|
H93-1019 |
phonetic transcription for accurate
|
speaker identification
|
. Being independent of the spoken
|
H93-1019 |
used for identification , the
|
speaker identification
|
accuracy is 100 % . This experimental
|
H93-1019 |
20 , 27 , 14 \ -RSB- with high
|
speaker identification
|
rates reported using subsets
|
H93-1019 |
been used in a few studies on
|
speaker identification
|
\ -LSB- l , 20 , 27 , 14 \ -RSB-
|
H93-1019 |
for accuracte text-independent
|
speaker identification
|
. ( EOS is End Of Sentence identification
|
H93-1019 |
. Performing text-independent
|
speaker identification
|
as before on the remaining 2
|
A97-1007 |
each turn , we store e.g. the
|
speaker identification
|
, the language of the contribution
|
H93-1013 |
for language identification ,
|
speaker identification
|
, speakersex identification ,
|
H93-1019 |
Experiments for text-dependent
|
speaker identification
|
using exactly the same models
|
D14-1051 |
identification ) . 4.2 From i-vector
|
speaker identification
|
to c-vector textual document
|