To improve the processing efficiency of a speech recognition device while maintaining speech recognition accuracy by adaptively changing the operation of a speech recognition processing section by a control signal corresponding to the quality of a speech data.
A frame sound quality estimation section 10 of the speech recognition device 100 outputs frame quality by evaluating sound quality of a speech digital signal for each frame, and an average sound quality estimating section 20 estimates a sound quality level from the frame sound quality of a plurality of frames. A speech recognition processing control section 30 outputs a control signal for controlling operation at speech recognition processing time to the speech recognition processing section based on the sound quality level. The speech recognition processing section 92' performs speech recognition processing based on the control signal.
YAMAGUCHI YOSHIKAZU
ASAMI TAICHI
JIN AKIO
MASATAKI HIROKAZU
TAKAHASHI SATOSHI
JPH08123467A | 1996-05-17 | |||
JP2004198832A | 2004-07-15 | |||
JP2000172291A | 2000-06-23 | |||
JP2009128490A | 2009-06-11 | |||
JPH10312194A | 1998-11-24 | |||
JPH09258765A | 1997-10-03 | |||
JP2000322080A | 2000-11-24 | |||
JP2001075594A | 2001-03-23 | |||
JP2000259181A | 2000-09-22 |
Taku Kusano
Yukio Nakamura