发明名称 SPEECH PRODUCTION ANALYSIS DEVICE, VOICE INTERACTION CONTROL DEVICE, METHOD, AND PROGRAM
摘要 PROBLEM TO BE SOLVED: To suitably understand the speech of a user even in a case where the result of voice recognition includes an unclear word due to user's incorrect reading.SOLUTION: A voice recognition part 11 voice-recognizes voice data indicating a speech of a user. An unclear word estimation part 12 estimates an unclear word that does not make sense in the speech of the user, from the voice recognition result according to a predetermined rule. A word analogizing part 13 analogizes a correct word for the unclear word based on an incorrectly read word and similarity, referring to an incorrectly read word list 15 in which a plurality of incorrectly read words are registered in the form of combinations of readings of elements into which words including a Chinese character is decomposed for one word or morpheme. An output part 14 outputs output data including the voice recognition result and the correct words.
申请公布号 JP2014145842(A) 申请公布日期 2014.08.14
申请号 JP20130013282 申请日期 2013.01.28
申请人 FUJITSU LTD 发明人 TAKAHASHI JUN
分类号 G10L15/20;G06F17/27;G06F17/28;G10L13/00;G10L15/22 主分类号 G10L15/20
代理机构 代理人
主权项
地址