发明名称 SPEECH RECOGNITION METHOD
摘要 PROBLEM TO BE SOLVED: To produce documents which are as close to verbatim transcripts of a source speech as possible.SOLUTION: A speech recognition method comprises: identifying a first phrase, the first phrase representing a first written form of a concept; identifying a two-way mapping between the first phrase and a second phrase, the second phrase representing a second written form of the concept; and replacing the first phrase with the second phrase.
申请公布号 JP2014029554(A) 申请公布日期 2014.02.13
申请号 JP20130211981 申请日期 2013.10.09
申请人 MULTIMODAL TECHNOLOGIES INC 发明人 KJELL SCHUBERT
分类号 G10L15/22;G06F3/16;G10L15/00 主分类号 G10L15/22
代理机构 代理人
主权项
地址