发明名称 |
SPEECH RECOGNIZER AND SPEECH RECOGNIZING METHOD |
摘要 |
<P>PROBLEM TO BE SOLVED: To perform a speech recognition robust to intrusion of non-stationary noise. <P>SOLUTION: A speech recognizer includes: a first air-conducted voice microphone 11 tightly mounted inside a body for picking up voice; a second air-conducted voice microphone 21 mounted outside the body for picking up the voice; a first decoder unit 16 for recognizing a first word string from the sound picked up by the first air-conducted voice microphone 11; a first word section extraction unit 19 for extracting a speech section corresponding to respective words constituting the first word string; a second decoder unit 26 for recognizing a second word string from the voice picked up by the second air-conducted voice microphone 21, on the speech section extracted by the first word section extraction unit 19; and a word string replacement unit 28 for replacing, with the second word string, the word string whose noise level is a prescribed value or less out of the first word string. <P>COPYRIGHT: (C)2012,JPO&INPIT |
申请公布号 |
JP2012145636(A) |
申请公布日期 |
2012.08.02 |
申请号 |
JP20110002027 |
申请日期 |
2011.01.07 |
申请人 |
MITSUBISHI ELECTRIC CORP;MITSUBISHI ELECTRIC BUILDING TECHNO SERVICE CO LTD |
发明人 |
GOI HIROTAKA;HANAZAWA TOSHIYUKI;IWASAKI TOMOHIRO;ABE SATOSHI |
分类号 |
G10L15/28;G10L15/20;H04R1/40 |
主分类号 |
G10L15/28 |
代理机构 |
|
代理人 |
|
主权项 |
|
地址 |
|