Presentation	2009-01-30 Speaker-direction Detection based on Emphasizing Particular Parts of Cross-correlation Function under Robot's Mechanical Noises Tsukasa Nunami, Takeshi Kawabata,
PDF Download Page	PDF download Page Link
Abstract(in Japanese)	(See Japanese page)
Abstract(in English)	A phoneme-cue based speaker-direction detection mechanism for toy/home robots is improved using the composite similarity method. In our previous work, the system searches for phoneme cues (spectral elements) from noisy speech with direct motor noises. Relatively good performance was archived when the phoneme cues were made from the speaker's own voice. However, the performance became worse using another person's phoneme cues. This paper describes a new method to improve the speaker-direction detection accuracy even if the speaker is unknown. Phoneme-cue models are constructed from multi-speaker vowels by utilizing the composite similarity method.
Keyword(in Japanese)	(See Japanese page)
Keyword(in English)
Paper #	SP2008-134
Date of Issue

Conference Information
Committee	SP
Conference Date	2009/1/22(1days)
Place (in Japanese)	(See Japanese page)
Place (in English)
Topics (in Japanese)	(See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To	Speech (SP)
Language	JPN
Title (in Japanese)	(See Japanese page)
Sub Title (in Japanese)	(See Japanese page)
Title (in English)	Speaker-direction Detection based on Emphasizing Particular Parts of Cross-correlation Function under Robot's Mechanical Noises
Sub Title (in English)
Keyword(1)
1st Author's Name	Tsukasa Nunami
1st Author's Affiliation	Kwansei Gakuin University()
2nd Author's Name	Takeshi Kawabata
2nd Author's Affiliation	Kwansei Gakuin University
Date	2009-01-30
Paper #	SP2008-134
Volume (vol)	vol.108
Number (no)	422
Page	pp.pp.-
#Pages	5
Date of Issue