Home >

news ヘルプ

論文・著書情報


タイトル
和文:Team Takoyaki submission for VoxCeleb Speaker Recognition Challenge 2020 
英文:Team Takoyaki submission for VoxCeleb Speaker Recognition Challenge 2020 
著者
和文: 石川 佳佑, 宇都 有昭, 岩野 公司, 篠田 浩一.  
英文: Keisuke Ishikawa, Kuniaki Uto, Koji Iwano, Koichi Shinoda.  
言語 English 
掲載誌/書名
和文: 
英文:Proc.The VoxSRC Workshop 2020 
巻, 号, ページ        
出版年月 2020年10月30日 
出版者
和文: 
英文:Visual Geometry Group 
会議名称
和文:The VoxSRC Workshop 2020 
英文:The VoxSRC Workshop 2020 
開催地
和文: 
英文: 
公式リンク http://www.robots.ox.ac.uk/~vgg/data/voxceleb/data_workshop_2020/takoyaki.pdf
 
アブストラクト This paper describes a system submitted to the VoxCelebSpeaker Recognition Challenge 2020 (VoxSRC2020). Previ-ously, Stafylakis et al. proposed a self-supervised learningmethod for x-vector-based speaker recognition. This methodtries to reconstruct the features of each input utterance fromits phoneme recognition result and the speaker features for thewhole utterance, but speaker features for each phoneme maybe different. In this paper, we propose a speaker recogni-tion method using phoneme-dependent speaker features. Thismethod concatenates the phone group label obtained fromphoneme recognition to the input features of each frame. Thenit generates frame-by-frame speaker features and use them forreconstruction. In the evaluation experiment, it achieved EERof 3.25%.

©2007 Institute of Science Tokyo All rights reserved.