Presentation	1999/12/20 Evaluation of Sub-Band Features Based on Information Criterion for Multi-Band Speech Recognition Takehiro NAKAJIMA, Shigeki OKAWA, Katsuhiko SHIRAI,
PDF Download Page	PDF download Page Link
Abstract(in Japanese)	(See Japanese page)
Abstract(in English)	This paper proposes an evaluation criterion based on information theory using sub-band features to improve the recognition performance of multi-band speech recognition. Particulmly, we discuss the correlation between the recognition performance and the conditional entropies of each sub-band feature derived from various freqency boundaries. As the result of phoneme recognition experiments, we observed an improvement in error reduction by 11.2% at 15dB SNR level by optimizing the sub-band features and we confirmed that our evaluation criterion is effective in noisy environments as well as in clean environment.
Keyword(in Japanese)	(See Japanese page)
Keyword(in English)
Paper #	NLC99-97
Date of Issue

Conference Information
Committee	NLC
Conference Date	1999/12/20(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)	Evaluation of Sub-Band Features Based on Information Criterion for Multi-Band Speech Recognition
Sub Title (in English)
Keyword(1)
1st Author's Name	Takehiro NAKAJIMA
1st Author's Affiliation	Dept. of Information and Computer Science, Waseda University()
2nd Author's Name	Shigeki OKAWA
2nd Author's Affiliation	Dept. of Network Science, Chiba Institute of Technology
3rd Author's Name	Katsuhiko SHIRAI
3rd Author's Affiliation	Dept. of Information and Computer Science, Waseda University
Date	1999/12/20
Paper #	NLC99-97
Volume (vol)	vol.99
Number (no)	523
Page	pp.pp.-
#Pages	6
Date of Issue