发明名称 VOICE RESPONSE SYSTEM, VOICE PROCESSING DEVICE, VOICE RESPONSE DEVICE, VOICE RESPONSE METHOD, AND VOICE PROCESSING PROGRAM
摘要 PROBLEM TO BE SOLVED: To increase accuracy for detecting a section in which speech is performed included in an audio signal.SOLUTION: The voice response system comprises an announcement reproduction unit, a background noise measurement unit, a voice detection unit, a voice recognition unit, and a response flow control unit. The announcement reproduction unit reproduces announcement for prompting speech of a speaker. The background noise measurement unit measures a background noise amount on the basis of an audio signal acquired in a silent section between announcement and speech of a speaker. The voice detection unit updates a voice detection threshold value on the basis of the background noise amount detected by the background noise measurement unit and detects a section in which a speaker performs speech by using the voice detection threshold value. The voice recognition unit performs recognition processing on the content of the speech by the speaker in the section detected by the voice detection unit. The response flow control unit selects announcement to be reproduced by the announcement reproduction unit on the basis of a recognition result of the voice recognition unit.
申请公布号 JP2015055790(A) 申请公布日期 2015.03.23
申请号 JP20130189709 申请日期 2013.09.12
申请人 TOSHIBA CORP 发明人 SAKAI SHIZUMARO
分类号 G10L15/22;G10L13/00;G10L13/02;G10L15/04 主分类号 G10L15/22
代理机构 代理人
主权项
地址