Home >

news Help

Publication Information


Title
Japanese: 
English:Speaker Separation in Multi-Channel Environment Using Deep Learning 
Author
Japanese: Liu Conggui, 井上 中順, 篠田浩一.  
English: Conggui Liu, Nakamasa Inoue, Koichi Shinoda.  
Language English 
Journal/Book name
Japanese:情報処理学会研究報告 
English:Technical Reports of IPSJ SLP 
Volume, Number, Page vol. 115    no. 11    pp. 1-6
Published date Feb. 18, 2017 
Publisher
Japanese:情報処理学会 
English:Information Processing Society of Japan 
Conference name
Japanese:第115回SLP研究発表会 
English:The 115th National Convention of IPSJ 
Conference site
Japanese:香川県仲多度郡琴平町977-1 
English: 
File
Official URL http://www.ipsj.or.jp/kenkyukai/event/slp115.html
 
Abstract This paper addresses multi-channel speaker separation based on a deep delay-and-subtraction beamformer. Deep neural network(DNN) is first applied to estimate the delay time between speakers and microphones , and then speakers’ speech is recovered from mixed signals by using a delay-and-subtraction algorithm. We evaluated our method by using simulated data made from WSJCAM0 database. The proposed method achieved high precision source localization, and about 62% relative improvement on word error rate (WER) over a delay-and-sum (DS) beamformer.

©2007 Tokyo Institute of Technology All rights reserved.