SEARCHING DEVICE, SEARCHING METHOD, AND PROGRAM

• US 20130006629A1
• Filed: 12/02/2010
• Published: 01/03/2013
• Est. Priority Date: 12/04/2009
• Status: Active Grant

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Honda, Hitoshi, Maeda, Yoshinori, Asakawa, Satoshi
• Granted Patent: US 9,817,889 B2
• Time in Patent Office: Days
• US Class Current: 704/236
• CPC Class Codes: G06F 16/3343 using phonetics
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/025 Phonemes, fenemes or fenones being the recognition units

Abstract

The present invention relates to a searching device, searching method, and program whereby searching for a word string corresponding to input voice can be performed in a robust manner.

A voice recognition unit 11 subjects an input voice to voice recognition. A matching unit 16 performs matching, for each of multiple word strings for search results which are word strings that are to be search results for word strings corresponding to the input voice, of a pronunciation symbol string for search results, which is an array of pronunciation symbols expressing pronunciation of the word string search result, and a recognition result pronunciation symbol string which is an array of pronunciation symbols expressing pronunciation of the voice recognition results of the input voice. An output unit 17 outputs a search result word string which is the result of searching the word strings corresponding to the input voice from the multiple word strings for search results, based on the matching results of the pronunciation symbol string for search results and the recognition result pronunciation symbol string. The present invention can be applied in the case of performing voice searching, for example.

Claims

1. A searching device comprising:

a voice recognition unit to subject input voice to voice recognition;

a matching unit to perform matching, regarding each of a plurality of word strings for search results which are word strings that are to be search results for word strings corresponding to said input voice,

of a pronunciation symbol string for search results, which is an array of pronunciation symbols expressing pronunciation of said word string search result, and

a recognition result pronunciation symbol string which is an array of pronunciation symbols expressing pronunciation of the voice recognition results of said input voice; and

an output unit to output a search result word string which is the result of searching the word strings corresponding to said input voice from said plurality of word strings for search results, based on the matching results of said pronunciation symbol string for search results and said recognition result pronunciation symbol string.

2. The searching device according to claim 1, wherein said pronunciation symbol is a symbol expressing the pronunciation of a syllable or a phoneme;

and wherein said matching unit,

in the matching of said pronunciation symbol string for search results and said recognition result pronunciation symbol string,

for a vector for search results which is a vector expressing said pronunciation symbol string for search results and a recognition result vector which is a vector expressing said recognition result pronunciation symbol string,

finds a correction distance wherein the cosine distance of a vector-space method has been corrected, so as to reduce influence from the difference in lengths between said pronunciation symbol string for search results and said recognition result pronunciation symbol string.

3. The searching device according to claim 2, further comprising;

a pronunciation symbol converting unit to convert the voice recognition results of said input voice into said recognition result pronunciation symbol string.

4. The searching device according to claim 3, wherein said pronunciation symbol converting unit further converts said word string for search results into said pronunciation symbol string for search results.

5. The searching device according to claim 2, wherein, in computations to find said cosine distance, said matching unit finds said correction distance by using the square root of a value that multiples the size of said vector for search results and the size of said recognition result vector, instead of the size of said vector for search results.

6. The searching device according to claim 2, wherein said in computations to find said cosine distance, said matching unit finds said correction distance by using the size of said recognition result vector.

7. A searching method that includes the steps of a searching device which searches for a word string corresponding to an input voice:

subjecting said input voice to voice recognition;

performing matching regarding each of a plurality of word strings for search results which are word strings that are to be search results for word strings corresponding to said input voice,

of a pronunciation symbol string for search results, which is an array of pronunciation symbols expressing pronunciation of said word string search result, and

a recognition result pronunciation symbol string which is an array of pronunciation symbols expressing pronunciation of the voice recognition results of said input voice; and

outputting a search result word string, which is the result of searching the word strings corresponding to said input voice from said plurality of word strings for search results, based on the matching results of said pronunciation symbol string for search results and said recognition result pronunciation symbol string.

8. A program causing a computer to function as:

a voice recognition unit to subject input voice to voice recognition;

a matching unit to perform matching regarding each of a plurality of word strings for search results which are word strings that are to be search results for word strings corresponding to said input voice,

of a pronunciation symbol string for search results, which is an array of pronunciation symbols expressing pronunciation of said word string search result, and

a recognition result pronunciation symbol string which is an array of pronunciation symbols expressing pronunciation of the voice recognition results of said input voice; and

an output unit to output a search result word string which is the result of searching the word strings corresponding to said input voice from said plurality of word strings for search results, based on the matching results of said pronunciation symbol string for search results and said recognition result pronunciation symbol string.