Home >

news Help

Publication Information


Title
Japanese: 
English:Speaker Diarization Using Multi-Modal i-vectors 
Author
Japanese: 西 史人, 井上 中順, 篠田 浩一.  
English: Fumito Nishi, Nakamasa Inoue, Koichi Shinoda.  
Language English 
Journal/Book name
Japanese: 
English:Proc. International Technical Conference on Circuits/Systems Computers and Communications (ITC-CSCC) 
Volume, Number, Page         pp. 27-30
Published date Jun 29, 2015 
Publisher
Japanese: 
English: 
Conference name
Japanese: 
English:The 30th International Technical Conference on Circuits/Systems, Computers and Communications (ITC-CSCC) 2015 
Conference site
Japanese:ソウル 
English:Seoul 
Abstract We propose multi-modal i-vectors, which extend the audio i-vector framework for speaker verification to a multi-modal speaker diarization in movies. In addition to the audio i-vector, which represents a speech utterance in an audio stream by a low-dimensional vector, we extract a visual i-vector from faces in a video segment. The audio and visual i-vectors are concatenated as a multi-modal i-vector clustered in an unsupervised way. We evaluate our method on the Hannah movie dataset. Our experiments show that diarization error rate is improved from 68.3% to 65.5% compared with audio stream only.

©2007 Tokyo Institute of Technology All rights reserved.