Presentation 2002/12/12
モーラ情報を用いた単語音声認識の検討
Takahiro SEO, Jin'ichi MURAKAMI, Satoru IKEHARA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Recently, in word speech recognition, cepstram is normally used as parameters. And, pitch information is not normally used because it is difficult to extract pitch frequency steady. However, in recent reserches, it is known that there is a relation between pitch frequency and mora information. Then in this paper, we performed word speech recognition using mora information in stead of pitch and examined its validity. In the results, by using mora information, error rate of word speech recognition could decrease for many speeker. And comparing the results of word speech recognition using mora information with using begining and ending of a word or using triphone, the recognition rate using mora information was higher. And the efficiency of this way was confirmed.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Word speech recognition / Mora length / Mora position / Begining and ending of a word / Triphone
Paper # NLC2002-53
Date of Issue

Conference Information
Committee NLC
Conference Date 2002/12/12(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 Natural Language Understanding and Models of Communication (NLC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English)
Sub Title (in English)
Keyword(1) Word speech recognition
Keyword(2) Mora length
Keyword(3) Mora position
Keyword(4) Begining and ending of a word
Keyword(5) Triphone
1st Author's Name Takahiro SEO
1st Author's Affiliation Faculty of Engineering, Tottori university()
2nd Author's Name Jin'ichi MURAKAMI
2nd Author's Affiliation Faculty of Engineering, Tottori university
3rd Author's Name Satoru IKEHARA
3rd Author's Affiliation Faculty of Engineering, Tottori university
Date 2002/12/12
Paper # NLC2002-53
Volume (vol) vol.102
Number (no) 527
Page pp.pp.-
#Pages 7
Date of Issue