1997
DOI: 10.1016/s0167-8655(97)00068-8
|View full text |Cite
|
Sign up to set email alerts
|

Impostor cohort selection for score normalisation in speaker verification

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
12
0

Year Published

2002
2002
2018
2018

Publication Types

Select...
3
3

Relationship

0
6

Authors

Journals

citations
Cited by 21 publications
(12 citation statements)
references
References 12 publications
0
12
0
Order By: Relevance
“…In previous work, 31 male speakers of the same age group (20-29 years) saying the words one to nine and zero were used (Finan et al 1997). Male speakers alone were used as this represents a generally harder identification problem than using a mixed male/female set.…”
Section: Results Of Wideband Recognitionmentioning
confidence: 99%
See 2 more Smart Citations
“…In previous work, 31 male speakers of the same age group (20-29 years) saying the words one to nine and zero were used (Finan et al 1997). Male speakers alone were used as this represents a generally harder identification problem than using a mixed male/female set.…”
Section: Results Of Wideband Recognitionmentioning
confidence: 99%
“…For some of the tests reported here, impostor cohort normalization (ICN) was used. The ICN derived scores have better equal error rates than the un-normalized scores (Finan et al 1997).…”
Section: Impostor Cohort Normalizationmentioning
confidence: 93%
See 1 more Smart Citation
“…In vector quantization (VQ) based speaker recognition [3,8,9,10,16], each speaker (or class) is presented by a codebook which approximates his/her data density by a small number of representative code vectors. Different regions (clusters) in the feature space represent acoustically different units.…”
Section: Introductionmentioning
confidence: 99%
“…The second discrimination paradigm, score normalization, is used in the decision phase. For instance, matching scores of the client speaker in speaker verification can be normalized against matching scores obtained from a cohort set [3].…”
Section: Introductionmentioning
confidence: 99%