Title:
SPEECH SEARCH DEVICE, COMPUTER-READABLE STORAGE MEDIUM, AND AUDIO SEARCH METHOD
Document Type and Number:
WIPO Patent Application WO/2014/033855
Kind Code:
A1
Abstract:
Provided is a speech search device which searches first speech data which is to be searched for a portion corresponding to a keyword which is inputted by a user. A speech search device, using second speech data, generates a sound model which denotes sound characteristics and a language model which denotes language characteristics. The speech search device converts the second audio data to a first sub-word string, converts a presumed keyword to a second sub-word string, computes a misrecognition trend with respect to the first sub-word string and the second sub-word string, converts first speech data to a third sub-word string, and converts a keyword to a fourth sub-word string. The speech search device searches the first sound data for a portion corresponding to the keyword as a search candidate. On the basis of the misrecognition trend, the speech search device computes a score on the basis of a sub-word score with respect to the third sub-word string with respect to the fourth sub-word string of the search candidate which is searched by a candidate search unit, and outputs a search result which includes the score and the search candidate corresponding to the score.
Inventors:
TAKEDA RYU (JP)
KANDA NAOYUKI (JP)
OBUCHI YASUNARI (JP)
SUMIYOSHI TAKASHI (JP)
KANDA NAOYUKI (JP)
OBUCHI YASUNARI (JP)
SUMIYOSHI TAKASHI (JP)
Application Number:
PCT/JP2012/071850
Publication Date:
March 06, 2014
Filing Date:
August 29, 2012
Export Citation:
Assignee:
HITACHI LTD (JP)
TAKEDA RYU (JP)
KANDA NAOYUKI (JP)
OBUCHI YASUNARI (JP)
SUMIYOSHI TAKASHI (JP)
TAKEDA RYU (JP)
KANDA NAOYUKI (JP)
OBUCHI YASUNARI (JP)
SUMIYOSHI TAKASHI (JP)
International Classes:
G06F17/30; G10L15/00; G10L15/18
Foreign References:
JP2010277036A | 2010-12-09 | |||
JP2011197410A | 2011-10-06 | |||
JP2010267012A | 2010-11-25 | |||
JP2005257954A | 2005-09-22 | |||
JP2009128508A | 2009-06-11 | |||
JP2009216986A | 2009-09-24 | |||
JP2011175046A | 2011-09-08 |
Other References:
NAOYUKI KANDA ET AL.: "Spoken Term Detection from Large Scale Speech Database Using Multistage Rescoring", THE TRANSACTIONS OF THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS, vol. J95-D, no. 4, 1 April 2012 (2012-04-01), pages 969 - 981
NAOYUKI KANDA ET AL.: "Evaluation of multistage rescoring strategy for open-vocabulary spoken term detection", DAI 2 KAI PROCEEDINGS OF THE SPOKEN DOCUMENT PROCESSING WORKSHOP, 1 March 2008 (2008-03-01), pages 73 - 78
NAOYUKI KANDA ET AL.: "Evaluation of multistage rescoring strategy for open-vocabulary spoken term detection", DAI 2 KAI PROCEEDINGS OF THE SPOKEN DOCUMENT PROCESSING WORKSHOP, 1 March 2008 (2008-03-01), pages 73 - 78
Attorney, Agent or Firm:
FUJII, MASAHIRO (JP)
Fujii Masahiro (JP)
Fujii Masahiro (JP)
Download PDF: