Audio Visual Speaker Verification based on Hybrid Fusion of Cross Modal Features

Girija Chetty, Michael Wagner

Research output: A Conference proceeding or a Chapter in BookConference contribution

6 Citations (Scopus)

Abstract

In this paper, we propose hybrid fusion of audio and explicit correlation features for speaker identity verification applications. Experiments were performed with the GMM based speaker models with a hybrid fusion technique involving late fusion of explicit cross-modal fusion features, with implicit eigen lip and audio MFCC features. An evaluation of the system performance with different gender specific datasets from controlled VidTIMIT data base and opportunistic UCBN database shows a significant performance improvement
Original languageEnglish
Title of host publicationPattern Recognition and Machine Intelligence Proceedings
EditorsA Ghosh, R.De
Place of PublicationGerman
PublisherSpringer
Pages469-478
Number of pages10
ISBN (Electronic)9783540770466
ISBN (Print)9783540770459
DOIs
Publication statusPublished - 2007
EventPattern Recognition and Machine Intelligence (PReMI) - , India
Duration: 18 Dec 200722 Dec 2007

Conference

ConferencePattern Recognition and Machine Intelligence (PReMI)
CountryIndia
Period18/12/0722/12/07

Fingerprint

Fusion reactions
Experiments

Cite this

Chetty, G., & Wagner, M. (2007). Audio Visual Speaker Verification based on Hybrid Fusion of Cross Modal Features. In A. Ghosh, & R.De (Eds.), Pattern Recognition and Machine Intelligence Proceedings (pp. 469-478). German: Springer. https://doi.org/10.1007/978-3-540-77046-6_58
Chetty, Girija ; Wagner, Michael. / Audio Visual Speaker Verification based on Hybrid Fusion of Cross Modal Features. Pattern Recognition and Machine Intelligence Proceedings. editor / A Ghosh ; R.De. German : Springer, 2007. pp. 469-478
@inproceedings{2638897bdf2547b8b60b8f0026e8096e,
title = "Audio Visual Speaker Verification based on Hybrid Fusion of Cross Modal Features",
abstract = "In this paper, we propose hybrid fusion of audio and explicit correlation features for speaker identity verification applications. Experiments were performed with the GMM based speaker models with a hybrid fusion technique involving late fusion of explicit cross-modal fusion features, with implicit eigen lip and audio MFCC features. An evaluation of the system performance with different gender specific datasets from controlled VidTIMIT data base and opportunistic UCBN database shows a significant performance improvement",
author = "Girija Chetty and Michael Wagner",
year = "2007",
doi = "10.1007/978-3-540-77046-6_58",
language = "English",
isbn = "9783540770459",
pages = "469--478",
editor = "A Ghosh and R.De",
booktitle = "Pattern Recognition and Machine Intelligence Proceedings",
publisher = "Springer",
address = "Netherlands",

}

Chetty, G & Wagner, M 2007, Audio Visual Speaker Verification based on Hybrid Fusion of Cross Modal Features. in A Ghosh & R.De (eds), Pattern Recognition and Machine Intelligence Proceedings. Springer, German, pp. 469-478, Pattern Recognition and Machine Intelligence (PReMI), India, 18/12/07. https://doi.org/10.1007/978-3-540-77046-6_58

Audio Visual Speaker Verification based on Hybrid Fusion of Cross Modal Features. / Chetty, Girija; Wagner, Michael.

Pattern Recognition and Machine Intelligence Proceedings. ed. / A Ghosh; R.De. German : Springer, 2007. p. 469-478.

Research output: A Conference proceeding or a Chapter in BookConference contribution

TY - GEN

T1 - Audio Visual Speaker Verification based on Hybrid Fusion of Cross Modal Features

AU - Chetty, Girija

AU - Wagner, Michael

PY - 2007

Y1 - 2007

N2 - In this paper, we propose hybrid fusion of audio and explicit correlation features for speaker identity verification applications. Experiments were performed with the GMM based speaker models with a hybrid fusion technique involving late fusion of explicit cross-modal fusion features, with implicit eigen lip and audio MFCC features. An evaluation of the system performance with different gender specific datasets from controlled VidTIMIT data base and opportunistic UCBN database shows a significant performance improvement

AB - In this paper, we propose hybrid fusion of audio and explicit correlation features for speaker identity verification applications. Experiments were performed with the GMM based speaker models with a hybrid fusion technique involving late fusion of explicit cross-modal fusion features, with implicit eigen lip and audio MFCC features. An evaluation of the system performance with different gender specific datasets from controlled VidTIMIT data base and opportunistic UCBN database shows a significant performance improvement

U2 - 10.1007/978-3-540-77046-6_58

DO - 10.1007/978-3-540-77046-6_58

M3 - Conference contribution

SN - 9783540770459

SP - 469

EP - 478

BT - Pattern Recognition and Machine Intelligence Proceedings

A2 - Ghosh, A

A2 - R.De, null

PB - Springer

CY - German

ER -

Chetty G, Wagner M. Audio Visual Speaker Verification based on Hybrid Fusion of Cross Modal Features. In Ghosh A, R.De, editors, Pattern Recognition and Machine Intelligence Proceedings. German: Springer. 2007. p. 469-478 https://doi.org/10.1007/978-3-540-77046-6_58