Publication:
Comparison of the impact of some Minkowski metrics on VQ/GMM based speaker recognition

dc.contributor.buuauthorHanilci, Cemal
dc.contributor.buuauthorErtaş, Figen
dc.contributor.departmentMühendislik Fakültesi
dc.contributor.departmentElektrik Elektronik Mühendisliği Bölümü
dc.contributor.researcheridS-4967-2016
dc.contributor.researcheridAAH-4188-2021
dc.contributor.scopusid35781455400
dc.contributor.scopusid24724154500
dc.date.accessioned2022-01-05T07:27:45Z
dc.date.available2022-01-05T07:27:45Z
dc.date.issued2011-01
dc.description.abstractThis paper evaluates the impact of three special forms of the Minkowski metric (Euclidean, City Block, and Chebychev distances) on the performance of the conventional vector quantization (VQ) and Gaussian mixture model (GMM) based closed-set text-independent speaker recognition systems, in terms of recognition rate and confidence on decisions. For the VQ based system, evaluations are carried out using the two most common clustering algorithms, LBG and K-means, and it is revealed which clustering algorithm and distance pair should be used to exploit the best attribute of both to achieve the best recognition rate for a given codebook size. In the case of GMM based system, we introduce the metrics into the GMM using a concatenation of the LBG and K-means algorithms in estimating the initial mean vectors, to which the system performance is sensitive, and explore their impact on system performance. We also make comparison of results obtained from evaluations on clean speech (TIMIT) and telephone speech databases (NTIMIT and NIST2001) with the modern classifiers VQ-UBM and GMM-UBM. It is found that there are cases where conventional VQ based system outperforms the modern systems. Moreover, the impact of distance metrics on the performance of the conventional and modern systems depends on the recognition task imposed (verification/identification).
dc.identifier.citationHanilci, C. vd. (2011). "Comparison of the impact of some Minkowski metrics on VQ/GMM based speaker recognition". Computers and Electrical Engineering, 37(1), 41-56.
dc.identifier.endpage56
dc.identifier.issn0045-7906
dc.identifier.issn1879-0755
dc.identifier.issue1
dc.identifier.scopus2-s2.0-79251600402
dc.identifier.startpage41
dc.identifier.urihttps://doi.org/10.1016/j.compeleceng.2010.08.001
dc.identifier.urihttps://dl.acm.org/doi/abs/10.1016/j.compeleceng.2010.08.001
dc.identifier.urihttp://hdl.handle.net/11452/23860
dc.identifier.volume37
dc.identifier.wos000287560300004
dc.indexed.wosSCIE
dc.language.isoen
dc.publisherPergamon-Elsevier Science
dc.relation.journalComputers and Electrical Engineering
dc.relation.publicationcategoryMakale - Uluslararası Hakemli Dergi
dc.rightsinfo:eu-repo/semantics/closedAccess
dc.subjectComputer science
dc.subjectEngineering
dc.subjectIdentification
dc.subjectAlgorithm
dc.subjectCharacter recognition
dc.subjectSpeech recognition
dc.subjectVector quantization
dc.subjectCity block
dc.subjectClean speech
dc.subjectCodebooks
dc.subjectDistance metrics
dc.subjectEuclidean
dc.subjectGaussian Mixture Model
dc.subjectK-means
dc.subjectk-Means algorithm
dc.subjectMean vector
dc.subjectMinkowski
dc.subjectMinkowski metrics
dc.subjectRecognition rates
dc.subjectSpeaker recognition
dc.subjectSpeaker recognition system
dc.subjectSpecial forms
dc.subjectTelephone speech
dc.subjectClustering algorithms
dc.subject.scopusSpeaker Verification; Language Recognition; Utterance
dc.subject.wosComputer science, hardware & architecture
dc.subject.wosComputer science, interdisciplinary applications
dc.subject.wosEngineering, electrical & electronic
dc.titleComparison of the impact of some Minkowski metrics on VQ/GMM based speaker recognition
dc.typeArticle
dc.wos.quartileQ3
dspace.entity.typePublication
local.contributor.departmentMühendislik Fakültesi/Elektrik Elektronik Mühendisliği Bölümü
local.indexed.atScopus
local.indexed.atWOS

Files

License bundle

Now showing 1 - 1 of 1
Placeholder
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed upon to submission
Description: