PROBLEM TO BE SOLVED: To search a part where a voice pronouncing a keyword is recorded from voice data without omission.;SOLUTION: A voice data search device includes: a phoneme table for registering at least a subset including phonemes whose language sounds are similar in the set of phonemes to be used for the voice expression of language; an acquisition part for acquiring the array of the phonemes included in a designated keyword; a generation part for displacing respective phonemes included in the keyword with the other phonemes in the subset including the phonemes to generate at least one derived keyword the array of whose phonemes is different; and a search part for searching a part in which similar voices which are similar to the keyword and the derived keyword are recorded as a similar part from the voice data.;COPYRIGHT: (C)2014,JPO&INPIT
展开▼