Dialogue supporting apparatus

• US 20050283365A1
• Filed: 06/27/2005
• Published: 12/22/2005
• Est. Priority Date: 04/12/2004
• Status: Active Grant

• Current Assignee: Panasonic Intellectual Property Corporation of America (Panasonic Holdings Corporation)
• Original Assignee: Panasonic Corporation (Panasonic Holdings Corporation)
• Inventors: Nambu, Taro, Mizutani, Kenji, Okimoto, Yoshiyuki
• Granted Patent: US 7,650,283 B2
• Time in Patent Office: Days
• US Class Current: 704/257
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/088 Word spotting

Abstract

The present invention provides a dialogue supporting apparatus that can easily select a desired sample sentence from among candidate-sample sentences corresponding to inputted speech. The dialogue supporting apparatus includes: a speech recognition unit which performs continuous speech recognition of the inputted speech; a database unit having a sample sentence database which holds the correspondence of sample sentences of a source language and a target language; a sample sentence selection unit which selects one or more sample sentences from within the sample sentence database, according to a speech recognition result or operation of a GUI unit; a sample sentence comparison unit which (i) compares the one or more sample sentences selected by the sample sentence selection unit and the speech recognition result, (ii) calculates word scores from an appearance location of the words, and (iii) derives a display parameter for each word of each sample sentence, based on the word scores; and the GUI unit which performs the display of a sample sentence based on the display parameter.

Claims

1. A dialogue supporting apparatus which recognizes speech inputted in a source language, and presents a corresponding translation in a target language, depending on a result of the speech recognition, said dialogue supporting apparatus comprising:

a sample sentence search unit operable to search out a sample sentence in the source language based on a keyword included in the speech recognition result;

a sample sentence comparison unit operable to compare the speech recognition result and the searched-out sample sentence; and

a sample sentence display control unit operable to display the searched-out sample sentence, and to highlight the keyword within the sample sentence based on the comparison of the speech recognition result and the sample sentence.

2. The dialogue supporting apparatus according to claim 1,

wherein said sample sentence comparison unit is operable to derive a keyword correspondence degree of the keyword by comparing a location within the sample sentence, of the keyword included in the speech recognition result and a location of the keyword within the speech recognition result, and

said sample sentence display control unit is operable to highlight the keyword within the sample sentence based on the keyword correspondence degree.

3. The dialogue supporting apparatus according to claim 2,

wherein said sample sentence comparison unit is operable to derive, based on the keyword correspondence degree, a display parameter for highlighting the keyword, and

said sample sentence display control unit is operable to highlight the keyword within the sample sentence based on the display parameter.

4. The dialogue supporting apparatus according to claim 2,

wherein said sample sentence comparison unit is operable to derive the location within the sample sentence, of the keyword included in the speech recognition result and the location of the keyword within the speech recognition result, using an appearance location in a character string of one of Japanese phonetic inscription and pronunciation symbols.

5. The dialogue supporting apparatus according to claim 4,

wherein said sample sentence comparison unit is operable to change the number of characters in one of the Japanese phonetic inscription and pronunciation symbols depending on a character type, and to count the changed number of characters, in deriving the location within the sample sentence, of the keyword included in the speech recognition result and the location of the keyword within the speech recognition result, using the appearance location in a character string of one of Japanese phonetic inscriptions and pronunciation symbols.

6. The dialogue supporting apparatus according to claim 5,

wherein said sample sentence comparison unit is operable to change the number of characters depending on whether the character type is any of a short sound, a prolonged sound, and a choked sound, of the Japanese phonetic inscriptions, and to count the changed number of characters, in the case where the source language is Japanese.

7. The dialogue supporting apparatus according to claim 5,

wherein said sample sentence comparison unit is operable to change the number of characters depending on whether the character type is any of a vowel, a long vowel, and a consonant, of the pronunciation symbols, and to count the changed number of characters, in the case where the source language is English.

8. The dialogue supporting apparatus according to claim 4,

wherein said sample sentence comparison unit is operable to adopt as the keyword correspondence degree, the difference between the location within the sample sentence, of the keyword included in the speech recognition result and the location of the keyword within the speech recognition result.

9. The dialogue supporting apparatus according to claim 4,

wherein said sample sentence comparison unit is operable to adopt as the keyword correspondence degree, a normalized value of the difference between the location within the sample sentence, of the keyword included in the speech recognition result and the location of the keyword within the speech recognition result.

10. The dialogue supporting apparatus according to claim 2,

wherein said sample sentence comparison unit is operable to derive a sentence correspondence degree for the sample sentence based on a keyword correspondence degree of each keyword included in the sample sentence.

11. The dialogue supporting apparatus according to claim 10,

wherein said sample sentence comparison unit is operable to derive as the sentence correspondence degree, the number of the keywords in the sample sentence, having a keyword correspondence degree that is not lower than a predetermined value.

12. The dialogue supporting apparatus according to claim 11,

wherein said sample sentence display control unit is operable to determine, when displaying a plurality of sample sentences, a display order of the sample sentences based on the sentence correspondence degree.

13. The dialogue supporting apparatus according to claim 12,

wherein said sample sentence display control unit is operable to display the sample sentences in declining order of sentence correspondence degree, when displaying the plurality of sample sentences.

14. The dialogue supporting apparatus according to claim 11,

wherein said sample sentence display control unit is operable to highlight an entire sentence of a sample sentence having a sentence correspondence degree that is not lower than a predetermined value, when displaying a plurality of sample sentences.

15. The dialogue supporting apparatus according to claim 14,

wherein said sample sentence comparison unit is operable to derive, based on the sentence correspondence degree, a display parameter for highlighting an entire sentence of the sample sentence, and

said sample sentence display control unit is operable to highlight the entire sentence of the sample sentence based on the display parameter.

16. A sample sentence searching apparatus which searches out a sample sentence corresponding to an inputted sentence, said sample sentence searching apparatus comprising:

a sample sentence search unit operable to search out the sample sentence based on a keyword included in the sentence;

a sample sentence comparison unit operable to compare the sentence and the searched-out sample sentence; and

a sample sentence display control unit operable to display the searched-out sample sentence, and to highlight the keyword within the sample sentence based on the comparison of the sentence and the sample sentence.

17. A dialogue support method for recognizing speech inputted in a source language, and presenting a corresponding translation in a target language, depending on a result of the speech recognition, said dialogue support method comprising:

searching out a sample sentence in the source language based on a keyword included in the speech recognition result;

comparing the speech recognition result and the searched-out sample sentence; and

displaying the searched-out sample sentence, and highlighting the keyword within the sample sentence based on the comparison of the speech recognition result and the sample sentence.

18. A program for recognizing speech inputted in a source language, and presenting a corresponding translation in a target language, depending on a result of the speech recognition, said program causing a computer to execute:

searching out a sample sentence in the source language based on a keyword included in the speech recognition result;

comparing the speech recognition result and the searched-out sample sentence; and

displaying the searched-out sample sentence, and highlighting the keyword within the sample sentence based on the comparison of the speech recognition result and the sample sentence.