PROBLEM TO BE SOLVED: To detect a conversation aiming for communication, from voice data dissolved into morphemes.;SOLUTION: A voice data analyzing device 2 includes: utterance section segment means 2c for discriminating a non-sound generating section as a delimiter of utterance sections, and segmenting the utterance section, when the length of the non-sound generating section exceeds a predetermined time period, concerning voice data of a target user and others classified into morpheme sections and the non-sound generating sections; response detecting means 2d for detecting the morpheme section of a response; response speaker discriminating means 2f for discriminating whether a speaker of the morpheme section of the response is the target person or others, about each utterance section; and conversation identifying means 2h for identifying the utterance section as a conversation section, when both of the target user and others are included as a speaker of the morpheme section of the response in one utterance section.;COPYRIGHT: (C)2014,JPO&INPIT
展开▼
