Presentation 2009-12-17
On a System with High-Speed Processing and Low-Power Consumption for Speech Signal Processing
Yoshikazu MIYANAGA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this report, a new noise robust speech recognition system is introduced. As a narrow sense of speech recognition, a system can recognize an observed waveform and then represents its result as an output. Such speech recognition systems are widely developed and presented. However, many users expect that a desired speech recognition system can observe any sound including speech, automatically extract a speech sound which should be recognized and then correctly recognize its speech waveform. It must show high recognition rate even under noisy circumstances. In this report, the system can automatically search a target speech sound among any observed data and recognize it with over 90% correctness under the conditions of 75dB sound-level circumstances and 300msec echo. This system is also designed as LSI hardware. It is shown that the designed system can execute on real-time and with low power consumption.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Speech Recognition / Speech Feature Extraction / Speech Detection / Speech Rejection / Noise Robust Speech Processing
Paper # SIS2009-42
Date of Issue

Conference Information
Committee SIS
Conference Date 2009/12/10(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 Smart Info-Media Systems (SIS)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) On a System with High-Speed Processing and Low-Power Consumption for Speech Signal Processing
Sub Title (in English)
Keyword(1) Speech Recognition
Keyword(2) Speech Feature Extraction
Keyword(3) Speech Detection
Keyword(4) Speech Rejection
Keyword(5) Noise Robust Speech Processing
1st Author's Name Yoshikazu MIYANAGA
1st Author's Affiliation Graduate School of Information Science and Technology, Hokkaido University()
Date 2009-12-17
Paper # SIS2009-42
Volume (vol) vol.109
Number (no) 338
Page pp.pp.-
#Pages 6
Date of Issue