Updating standard patterns of words in a voice recognition dictionary

• US 8,438,027 B2
• Filed: 05/25/2006
• Issued: 05/07/2013
• Est. Priority Date: 05/27/2005
• Status: Expired due to Fees

• Current Assignee: Panasonic Intellectual Property Corporation of America (Panasonic Holdings Corporation)
• Original Assignee: Panasonic Corporation (Panasonic Holdings Corporation)
• Inventors: Teranishi, Toshiyuki, Hatano, Kouji
• Primary Examiner(s): PULLIAS, JESSE SCOTT
• Application Number:
  US11/915,613
• Publication Number:
  US 20090106027A1
• Time in Patent Office: 2,539 Days
• Field of Search: None
• US Class Current: 704/243
• CPC Class Codes: G10L 15/06 Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice G10L15/14 takes precedence
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering

Abstract

An object of the invention is to conveniently increase standard patterns registered in a voice recognition device to efficiently extend the amount of words that can be voice-recognized. New standard patterns are generated by modifying a part of an existing standard pattern. A pattern matching unit 16 of a modifying-part specifying unit 14 performs pattern matching process to specify a part to be modified in the existing standard pattern of a usage source. A standard pattern generating unit 18 generates the new standard patterns by cutting or deleting voice data of the modifying part of the usage-source standard pattern, substituting the voice data of the modifying part of the usage-source standard pattern for another voice data, or combining the voice data of the modifying part of the usage-source standard pattern with another voice data. A standard pattern database update unit 20 adds the new standard patterns to a standard pattern database 24.

Claims

1. A voice edition apparatus, of a mobile terminal, that uses existing voice data to generate another voice data, comprising:

a standard pattern database that is stored in the mobile terminal, the standard pattern database storing the existing voice data as a plurality of predefined standard patterns, each standard pattern representing a plurality of words in a sequence;

a modifying-part specifying unit that performs a matching process between usage-source voice data among the existing voice data and modifying-part voice data for specifying a modifying part of the usage-source voice data, and specifies at least a part of the usage-source voice data as the modifying part, wherein the usage-source voice data represents a plurality of words in a sequence; and

a voice data generating unit that uses the modifying-part voice data specified by the modifying-part specifying unit to generate another voice data as another standard pattern representing a new plurality of words in a sequence not present in the standard pattern database,

wherein a sequence of words within the usage-source voice data is identical to a sequence of words within the another standard pattern, the sequence of words within the usage-source voice data being a subset of said plurality of words in a sequence, and the sequence of words within the another standard pattern being a subset of said new plurality of words in a sequence not present in the standard pattern database, the another standard pattern including a further subset sequence of words that does not match the usage-source voice data,

wherein the standard pattern database is updated by storing the another voice data in the standard pattern database.

2. The voice edition apparatus of claim 1, wherein the voice data generating unit generates, as another voice data, voice data obtained by substituting the modifying-part voice data of the usage-source voice data specified by the modifying-part specifying unit with different voice data.

3. The voice edition apparatus of claim 2, wherein the different voice data used to generate another voice data comprises voice data that is input to the voice edition apparatus from the outside or voice data that is cut from the existing voice data.

4. The voice edition apparatus of claim 1, wherein the voice data generating unit generates, as another voice data, voice data obtained by combining the modifying-part voice data specified by the modifying-part specifying unit with different voice data.

5. The voice edition apparatus of claim 1, wherein the mobile terminal recognizes that a broadcast announcement matches one of the predefined standard patterns and, based thereon, activates a vibrator.

6. The voice edition apparatus of claim 1, wherein the mobile terminal recognizes that a broadcast announcement matches one of the predefined standard patterns and, based thereon, automatically turns itself off.

7. A voice recognition apparatus that uses one of the predefined standard patterns to generate the another standard pattern by means of the voice edition apparatus of claim 1,

wherein the modifying-part specifying unit performs a matching process between a usage-source standard pattern among the predefined standard patterns and modifying-part voice data for specifying a modifying-part of the usage-source standard pattern and specifies at least a part of the usage-source standard pattern as the modifying part;

and

wherein the voice data generating unit uses the modifying-part specifying unit to generate the another standard pattern.

8. The voice recognition apparatus of claim 7, further comprising a standard pattern database update unit that updates the standard pattern database by substituting the another standard pattern generated by the voice data generating unit for the usage-source standard pattern, or by adding the another standard pattern as a standard pattern that is different from the usage-source standard pattern.

9. The voice recognition apparatus of claim 7, wherein the modifying-part specifying unit performs a matching process between the voice data for extracting the usage-source standard pattern and the voice data having a plurality of existing standard patterns, and extracts, as the usage-source standard pattern, the predefined standard pattern containing a matched part between the voice data for extracting the usage-source standard pattern and the voice data having a plurality of existing standard patterns.

10. A voice edition method that uses existing voice data to generate another voice data, the method comprising:

recording a sound using a microphone;

generating modifying-part voice data based on the sound;

providing a standard pattern database storing existing voice data as a plurality of predefined standard patterns, each standard pattern representing a plurality of words in a sequence;

performing a match process between usage-source voice data among the existing voice data and the modifying-part voice data for specifying a modifying part of the usage-source voice data and specifying at least a part of the usage-source voice data as the modifying part,

using the specified modifying-part voice data to generate another voice data as another standard pattern representing a new plurality of words in a sequence not present in the standard pattern database, wherein a sequence of words within the usage-source voice data is identical to a sequence of words within the another standard pattern, the sequence of words within the usage-source voice data being a subset of a plurality of words in a sequence represented by the usage-source voice data, and the sequence of words within the another standard pattern being a subset of said new plurality of words in a sequence not present in the standard pattern database, the another standard pattern including a further subset sequence of words that does not match the usage-source voice data; and

updating the standard pattern database by storing the another voice data in the standard pattern database.

11. The voice edition method of claim 10, further comprising the steps of:

recognizing, by a mobile terminal, that a broadcast announcement matches one of the predefined standard patterns; and

activating a vibrator based on a result of the step of recognizing.

12. The voice edition method of claim 10, further comprising the steps of:

recognizing, by a mobile terminal, that a broadcast announcement matches one of the predefined standard patterns; and

the mobile terminal automatically turning itself off based on a result of the step of recognizing.

13. A voice edition program product embodied on a non-transitory computer readable medium which, when executed by a computer, cause the computer to execute the voice edition method of claim 10.

14. A voice recognition method that uses a predefined standard pattern to generate another standard pattern, the method comprising:

recording a sound using a microphone;

generating modifying-part voice data based on the sound;

providing a standard pattern database storing a plurality of predefined standard patterns, each standard pattern representing a plurality of words in a sequence;

performing a matching process between a usage-source standard pattern among the predefined standard patterns and the modifying-part voice data for specifying a modifying part of the usage-source standard pattern and specifying at least a part of the usage-source standard pattern as the modifying part;

using the specified modifying-part voice data to generate another standard pattern representing a new plurality of words in a sequence not present in the standard pattern database, wherein a sequence of words within the usage-source standard pattern is identical to a sequence of words within the another standard pattern, the sequence of words within the usage-source standard pattern being a subset of a plurality of words in a sequence represented by the usage-source standard pattern, and the sequence of words within the another standard pattern being a subset of said new plurality of words in a sequence not present in the standard pattern database, the another standard pattern including a further subset sequence of words that does not match the usage-source standard pattern; and

updating the standard pattern database by storing the another standard pattern in the standard pattern database.

15. The voice recognition method of claim 14, further comprising the steps of:

recognizing, by a mobile terminal, that a broadcast announcement matches one of the predefined standard patterns; and

activating a vibrator based on a result of the step of recognizing.

16. The voice recognition method of claim 14, further comprising the steps of:

recognizing, by a mobile terminal, that a broadcast announcement matches one of the predefined standard patterns; and

the mobile terminal automatically turning itself off based on a result of the step of recognizing.

17. A voice edition program product embodied on a non-transitory computer readable medium which, when executed by a computer, cause the computer to execute the voice recognition method of claim 14.