CASIA OpenIR  > 模式识别国家重点实验室  > 多媒体计算与图形学
Centroid-aware local discriminative metric learning in speaker verification
Sheng, Kekai1,2; Dong, Weiming1; Li, Wei3; Razik, Joseph4; Huang, Feiyue3; Hu, Baogang1; Weiming Dong
AbstractWe propose a new mechanism to pave the way for efficient learning against class-imbalance and improve representation of identity vector (i-vector) in automatic speaker verification (ASV). The insight is to effectively exploit the inherent structure within ASV corpus - centroid priori. In particular: (1) to ensure learning efficiency against class-imbalance, the centroid-aware balanced boosting sampling is proposed to collect balanced mini-batch; (2) to strengthen local discriminative modeling on the mini-batches, neighborhood component analysis (NCA) and magnet loss (MNL) are adopted in ASV-specific modifications. The integration creates adaptive NCA (AdaNCA) and linear MNL (LMNL). Numerical results show that LMNL is a competitive candidate for low-dimensional projection on i-vector (EER=3.84% on SRE2008, EER=1.81% on SRE2010), enjoying competitive edge over linear discriminant analysis (LDA). AdaNCA (EER=4.03% on SRE2008, EER=2.05% on SRE2010) also performs well. Furthermore, to facilitate the future study on boosting sampling, connections between boosting sampling, hinge loss and data augmentation have been established, which help understand the behavior of boosting sampling further. (C) 2017 Elsevier Ltd. All rights reserved.
KeywordText-independent Asv Centroid-aware Balanced Boosting Sampling Adaptive Neighborhood Component Analysis Linear Magnet
WOS HeadingsScience & Technology ; Technology
Indexed BySCI
Funding OrganizationNational Natural Science Foundation of China (NSFC)(61573348 ; Institute of Automation Chinese Academy of Sciences (CASIA)-Tencent Youtu Joint Research Project ; 61620106003 ; 61672520)
WOS Research AreaComputer Science ; Engineering
WOS SubjectComputer Science, Artificial Intelligence ; Engineering, Electrical & Electronic
WOS IDWOS:000411545400013
Citation statistics
Document Type期刊论文
Corresponding AuthorWeiming Dong
Affiliation1.Chinese Acad Sci, Inst Automat, LIAMA NLPR, Beijing, Peoples R China
2.Univ Chinese Acad Sci, Beijing, Peoples R China
3.Tencent Inc, Lab Youtu, Shanghai, Peoples R China
4.Univ Toulon & Var, Lab LSIS, Toulon, France
Recommended Citation
GB/T 7714
Sheng, Kekai,Dong, Weiming,Li, Wei,et al. Centroid-aware local discriminative metric learning in speaker verification[J]. PATTERN RECOGNITION,2017,72(72):176-185.
APA Sheng, Kekai.,Dong, Weiming.,Li, Wei.,Razik, Joseph.,Huang, Feiyue.,...&Weiming Dong.(2017).Centroid-aware local discriminative metric learning in speaker verification.PATTERN RECOGNITION,72(72),176-185.
MLA Sheng, Kekai,et al."Centroid-aware local discriminative metric learning in speaker verification".PATTERN RECOGNITION 72.72(2017):176-185.
Files in This Item: Download All
File Name/Size DocType Version Access License
Centroid-aware local(2013KB)期刊论文作者接受稿开放获取CC BY-NC-SAView Download
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Sheng, Kekai]'s Articles
[Dong, Weiming]'s Articles
[Li, Wei]'s Articles
Baidu academic
Similar articles in Baidu academic
[Sheng, Kekai]'s Articles
[Dong, Weiming]'s Articles
[Li, Wei]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Sheng, Kekai]'s Articles
[Dong, Weiming]'s Articles
[Li, Wei]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: Centroid-aware local discriminative metric learning in speaker verification.pdf
Format: Adobe PDF
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.