Cartridge-based, interactive speech recognition method with a response creation capability

• US 5,946,658 A
• Filed: 10/02/1998
• Issued: 08/31/1999
• Est. Priority Date: 08/21/1995
• Status: Expired due to Term

• Current Assignee: Seiko Epson Corporation (Seiko Group)
• Original Assignee: Seiko Epson Corporation (Seiko Group)
• Inventors: Miyazawa, Yasunaga, Inazumi, Mitsuhiro, Hasegawa, Hiroshi, Urano, Osamu, Edatsune, Isao
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): Smits, Talivaldis Ivars
• Application Number:
  US09/165,512
• Time in Patent Office: 333 Days
• Field of Search: 704/244, 704/251, 704/258, 704/275
• US Class Current: 704/275
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/0638 Interactive procedures
• CPC Class Codes: G10L 2015/088 Word spotting

Abstract

A technique for improving speech recognition in low-cost, speech interactive devices. This technique calls for selectively implementing a speaker-specific word enrollment and detection unit in parallel with a word detection unit to permit comprehension of spoken commands or messages when no recognizable words are found. Preferably, specific speaker detection will be based on the speaker'"'"'s own personal list of words or expression. Other facets include complementing non-specific pre-registered word characteristic information with individual, speaker-specific verbal characteristics to improve recognition in cases where the speaker has unusual speech mannerisms or accent and response alteration in which speaker-specification registration functions are leveraged to provide access and permit changes to a predefined responses table according to user needs and tastes. Also disclosed is the externalization and modularization of non-specific speaker recognition, action and response information to enhance adaptability of the speech recognizer without sacrificing product cost competitiveness or overall device responsiveness.

Claims

1. A method for performing interactive speech recognition processing, comprising the steps of:

receiving voice and translating the received voice into digital form;

generating characteristic voice data for the received digitized voice;

determining whether the characteristic voice data substantially matches standard characteristic voice information corresponding to pre-registered expressions and generating phrase identification data in response thereto, wherein the pre-registered expressions are stored as standard speech patterns capable of recognition in a removable cartridge releasably communicating with said phrase identification unit, said removable cartridge comprising a first memory to retain the standard speech patterns;

recognizing a meaning from the received voice based on the received phrase identification data and formulating an appropriate response corresponding to the recognized meaning;

enabling the creation of response data based on inputted information; and

generating synthesized audio corresponding to the appropriate response formulated in said recognizing and formulating step.

2. The method of claim 1, wherein said removable cartridge includes a second memory to retain conversation content data used to recognize the meaning from the received and recognized voice.

3. The method of claim 2, wherein said removable cartridge includes a third memory to retain response data used to formulate and synthesize the appropriate response to the received and recognized voice.

4. The method of claim 3, wherein said first, second and third cartridge memories reside within at least one ROM device.

5. The method of claim 3, wherein said first, second and third cartridge memories reside within at least one EEPROM device.

6. The method of claim 1, wherein said removable cartridge includes a second memory to retain response data used to formulate and synthesize the appropriate response to the received and recognized voice.

7. A method for performing interactive speech recognition processing, comprising the steps of:

receiving voice and translating the received voice into digital form;

generating characteristic voice data for the received digitized voice;

determining whether the characteristic voice data substantially matches standard characteristic voice information corresponding to pre-registered expressions and generating phrase identification data in response thereto;

recognizing a meaning from the received voice based on the received phrase identification data and conversation content information stored in a first memory of a removable cartridge releasably communicating therewith, and formulating an appropriate response corresponding to the recognized meaning;

enabling the creation of response data based on inputted information; and

generating synthesized audio corresponding to the appropriate response formulated in said recognizing and formulating step.

8. The method of claim 7, wherein said removable cartridge includes a second memory to retain response data used to formulate and synthesize the appropriate response to the received and recognized voice.

9. The method of claim 8, wherein said first and second cartridge memories reside within at least one ROM device.

10. The method of claim 8, wherein said first and second cartridge memories reside within at least one EEPROM device.