摘要 |
<P>PROBLEM TO BE SOLVED: To provide a voice data retrieval system capable of easily determining whether a search result is correct or incorrect. <P>SOLUTION: The voice data retrieval system includes: an input device 112 for inputting a keyword; a phoneme conversion unit 106 for converting the inputted the keyword into phoneme notation; a voice data search unit 105 which searches a part where the keyword is spoken in voice data on the basis of the keyword of phoneme notation; a comparison keyword generation unit 107 for generating a set of comparison keywords which may confuse a user hearing them on the basis of a phoneme confusion matrix 103 for each user; a search result presentation unit 110 which presents the search result from the voice data search unit 105 and the comparison keyword from the comparison keyword generation unit 107 to the user. <P>COPYRIGHT: (C)2013,JPO&INPIT |