Investigation of DNN based feature enhancement jointly trained with x-vectors for noise-robust speaker verification
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Yang, Joon-Young | - |
dc.contributor.author | Park, Kwan-Ho | - |
dc.contributor.author | Chang, Joon-Hyuk | - |
dc.contributor.author | Kim, Youngsam | - |
dc.contributor.author | Cho, Sangrae | - |
dc.date.accessioned | 2021-07-30T05:22:47Z | - |
dc.date.available | 2021-07-30T05:22:47Z | - |
dc.date.created | 2021-05-11 | - |
dc.date.issued | 2020-01 | - |
dc.identifier.uri | https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/4465 | - |
dc.description.abstract | In this paper, we investigate the deep neural network (DNN) based feature enhancement as the denoising frontend of the x-vector speaker verification framework in noisy environments. Firstly, the feature enhancement DNN (FE-DNN) learns the mapping function from the noisy to the clean corpora on the frame-level acoustic feature domain, and then the x-vector network (XvectorNet) is trained on top of the enhanced features. Finally, the separately trained FE-DNN and the XvectorNet are serially concatenated and jointly trained under the supervision of cross-entropy loss. In addition., we adopt the logistic margin softmax layer for training the XvectorNet in order to obtain more discriminative speaker embeddings. | - |
dc.language | 영어 | - |
dc.language.iso | en | - |
dc.publisher | Institute of Electrical and Electronics Engineers Inc. | - |
dc.title | Investigation of DNN based feature enhancement jointly trained with x-vectors for noise-robust speaker verification | - |
dc.type | Article | - |
dc.contributor.affiliatedAuthor | Chang, Joon-Hyuk | - |
dc.identifier.doi | 10.1109/ICEIC49074.2020.9051093 | - |
dc.identifier.scopusid | 2-s2.0-85083494899 | - |
dc.identifier.bibliographicCitation | 2020 International Conference on Electronics, Information, and Communication, ICEIC 2020, pp.1 - 5 | - |
dc.relation.isPartOf | 2020 International Conference on Electronics, Information, and Communication, ICEIC 2020 | - |
dc.citation.title | 2020 International Conference on Electronics, Information, and Communication, ICEIC 2020 | - |
dc.citation.startPage | 1 | - |
dc.citation.endPage | 5 | - |
dc.type.rims | ART | - |
dc.type.docType | Conference Paper | - |
dc.description.journalClass | 1 | - |
dc.description.isOpenAccess | N | - |
dc.description.journalRegisteredClass | scopus | - |
dc.subject.keywordPlus | Speech recognition | - |
dc.subject.keywordPlus | Acoustic features | - |
dc.subject.keywordPlus | Cross entropy | - |
dc.subject.keywordPlus | Feature enhancement | - |
dc.subject.keywordPlus | Mapping functions | - |
dc.subject.keywordPlus | Noise robust | - |
dc.subject.keywordPlus | Noisy environment | - |
dc.subject.keywordPlus | Speaker verification | - |
dc.subject.keywordPlus | Vector networks | - |
dc.subject.keywordPlus | Deep neural networks | - |
dc.subject.keywordAuthor | Deep speaker embedding | - |
dc.subject.keywordAuthor | Feature enhancement | - |
dc.subject.keywordAuthor | Joint training | - |
dc.subject.keywordAuthor | Speaker verification | - |
dc.identifier.url | https://ieeexplore.ieee.org/document/9051093 | - |
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.
222, Wangsimni-ro, Seongdong-gu, Seoul, 04763, Korea+82-2-2220-1365
COPYRIGHT © 2021 HANYANG UNIVERSITY.
Certain data included herein are derived from the © Web of Science of Clarivate Analytics. All rights reserved.
You may not copy or re-distribute this material in whole or in part without the prior written consent of Clarivate Analytics.