Title:
音声認識装置
Document Type and Number:
Japanese Patent JP4101365
Kind Code:
B2
Abstract:
PROBLEM TO BE SOLVED: To improve the speech recognition rate by changing contents of a message for requesting a speaker to reinput a speech according to the time difference between an utterance start point and an acceptance start point of time when the utterance start point of the speaker is precedent to the acceptance start point. SOLUTION: It is judged whether or not the start of utterance is earlier than a start sound (S1). In this case, the judgement is made from the difference between the time when the start sound 'beep' is generated and the time when a speech input part detects the speech of the speaker. When the start is much earlier, the speaker is warned to speak after hearing the start sound (S3) by using a word 'after making sure' as a message requesting revoicing. Then a message 'Speak more slowly after the sound 'beep' is reproduced (S4). Thus, the contents of the message to the speaker are changed according to the earliness of the start of utterance. The speaker can adjust the timing of the start of utterance.
Inventors:
Masahiro Kamiya
Toshitaka Yamato
Toshiaki Kusano
Kazuhiro Sakiyama
Hideki Kitao
Toshitaka Yamato
Toshiaki Kusano
Kazuhiro Sakiyama
Hideki Kitao
Application Number:
JP21077198A
Publication Date:
June 18, 2008
Filing Date:
July 27, 1998
Export Citation:
Assignee:
Fujitsu Ten Co., Ltd.
International Classes:
G10L15/22
Domestic Patent References:
| JP8335094A | ||||
| JP59170936A | ||||
| JP64048099A | ||||
| JP2014990A | ||||
| JP62294297A | ||||
| JP63259689A |