Speech recognition in consumer electronic products

• US 20020091513A1
• Filed: 01/16/2002
• Published: 07/11/2002
• Est. Priority Date: 12/06/1996
• Status: Active Grant

• Current Assignee: Sensory Incorporated
• Original Assignee: Sensory Incorporated
• Inventors: Mozer, Todd F., North, Thomas, Mozer, Forrest S.
• Granted Patent: US 6,665,639 B2
• Time in Patent Office: Days
• US Class Current: 704/220
• CPC Class Codes: G04G 21/06 using voice
• CPC Class Codes: G10L 15/285 Memory allocation or algorithm optimisation to reduce hardware requirements

Abstract

A method and apparatus are described that allow inexpensive speech recognition in applications where this capability is not otherwise feasible because of cost or technical reasons, or because of inconvenience to the user. A relatively simple speaker independent recognition algorithm, capable of recognizing a limited number of utterances at any one time, is associated with the base unit of an electronics product. To function, the product requires information from an external medium and this medium also provides the data required to recognize several sets of utterances pertinent to other information provided by the external medium.

Claims

1. In a watch wherein features of spoken utterances are analyzed by a programmable speaker-independent pattern recognition system to provide recognition results, a method comprising the steps of:

coupling said watch to a source external to said watch;

retrieving programming information for said pattern recognition system from said external source; and

programming said pattern recognition system with said retrieved programming information to recognize a first set of words, wherein said speaker-independent pattern recognition system provides a first set of recognition results responsive to said first set of words.

2. The method of claim 1, wherein said programming information is available in a plurality of languages, and wherein said retrieved programming information is in at least one of said plurality of languages.

3. The method of claim 1, wherein said retrieved programming information further comprises at least one voice prompt.

4. The method of claim 1, wherein said pattern recognition system initially recognizes an initial set of words in a first language, and wherein said first set of words is in a second language.

5. The method of claim 1, further comprising the steps of:

retrieving a second set of programming information for said pattern recognition system from said external source; and

programming said pattern recognition system with said retrieved second set of programming information to recognize a second set of words, wherein said speaker-independent pattern recognition system provides a second set of recognition results responsive to said second set of words.

6. The method of claim 1, wherein said retrieved programming information comprises at least two speaker-independent recognition sets.

7. The method of claim 6, wherein a first of said two speaker-independent recognition sets pertains to digits and a second of said two speaker-independent recognition sets pertains to time-setting commands.

8. A watch, comprising:

a programmable speaker-independent pattern recognition system;

a source of programming information external to said watch, said source of programming information comprising a time setting function for said watch;

an interface coupleable to said external source, wherein said interface retrieves said programming information from said external source and programs said programmable speaker-independent pattern recognition system to recognize a first set of words, wherein said watch adjusts a current time in response to said first set of words; and

a display that displays said current time.

9. The watch of claim 8, wherein said programming information is available from said external source in a plurality of languages.

10. The watch of claim 9, wherein said first set of words is in at least a first language of said plurality of languages.

11. The watch of claim 8, wherein said external source is selected from the group consisting of wireless networks, external storage media, telephone networks, CATV, Internet networks, and Intranet networks.

12. The watch of claim 8, wherein said programming information further comprises a plurality of voice prompts.

14. The method of claim 13, further comprising the step of operating said pattern recognition system as programmed in said programming step to generate at least a first recognition result responsive to input speech.

15. The method of claim 13, further comprising the steps of:

retrieving a second programming information for said pattern recognition system from a second source external to said base unit; and

reprogramming said pattern recognition system with said second programming information to recognize a second set of words.

16. The method of claim 13, wherein said external source is selected from the group consisting of wireless networks, external storage media, telephone networks, CATV, Internet networks, and Intranet networks.

17. The method of claim 13, wherein said external source stores an indication that said pattern recognition programming information is stored on said external source.

18. The method of claim 13, further comprising the step of verifying that pattern recognition programming information is available from said external source.

19. The method of claim 13, further comprising the step of displaying graphical information retrieved from said external source responsive to said first recognition result.

20. The method of claim 13, wherein said retrieved programming information includes information for synthesizing speech.

21. The method of claim 13, wherein said retrieved programming information includes information for generating graphical displays.

22. The method of claim 13, wherein said retrieved programming information includes neural network weights.

24. The apparatus of claim 23, wherein said apparatus is selected from the group consisting of televisions, computers, video game players, audio players, watches, and telephones.

25. The apparatus of claim 23, wherein said interface is selected from the group of interfaces consisting of wireless network interfaces, Internet interfaces, external storage media, telephone interfaces, and CATV interfaces.

26. The apparatus of claim 23, further comprising a display that displays graphical information in response to said recognition results.

27. The apparatus of claim 23, further comprising a speech synthesis system that generates speech in response to said recognition results.

13. In a base unit wherein features of spoken utterances are analyzed by a programmable speaker-independent pattern recognition system to provide recognition results, a method comprising the steps of:

retrieving programming information for said pattern recognition system from a source external to said base unit; and

programming said pattern recognition system with said retrieved programming information to recognize a first set of words, wherein said speaker-independent pattern recognition system provides a first set of recognition results responsive to said first set of words.

23. Apparatus for performing speaker independent speech recognition comprising:

a programmable pattern recognition system;

an interface that retrieves programming information for programming said programmable pattern recognition system to recognize a selected word vocabulary;

a source external to said apparatus that supplies said programming information to said interface for programming said programmable pattern recognition system;

a memory that stores said programming information for said programmable pattern recognition system to recognize a selected word vocabulary; and

a result generator for generating at least a first recognition result responsive to at least a portion of said selected word vocabulary.