Speech recognition system with display information

• US 8,364,487 B2
• Filed: 10/21/2008
• Issued: 01/29/2013
• Est. Priority Date: 10/21/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Ju, Yun-Cheng, Odell, Julian J.
• Primary Examiner(s): Chawan, Vijay B
• Application Number:
  US12/255,270
• Publication Number:
  US 20100100384A1
• Time in Patent Office: 1,561 Days
• Field of Search: 704/270, 704/251, 704/235, 704/8, 704/257, 704/277, 704/3, 704/9, 704/256, 704/10, 704/243, 704/275, 704/256.1, 704/260, 704/231, 704/258, 704/252, 704/271, 379/88.01, 434/185, 715/703, 381/2, 455/466
• US Class Current: 704/257
• CPC Class Codes: G10L 15/19 Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A language processing system may determine a display form of a spoken word by analyzing the spoken form using a language model that includes dictionary entries for display forms of homonyms. The homonyms may include trade names as well as given names and other phrases. The language processing system may receive spoken language and produce a display form of the language while displaying the proper form of the homonym. Such a system may be used in search systems where audio input is converted to a graphical display of a portion of the spoken input.

Claims

1. A method comprising:

receiving a spoken search request;

decoding said spoken search request into a sequence of words;

processing said sequence of words using a normalizing language model to determine a plurality of sets of normalized words representing said spoken search request, at least one of said words being a homonym;

said normalizing language model comprising homonym rules defining a plurality of display forms for a homonym; and

generating said plurality of sets of words in said display form.

2. The method of claim 1 further comprising:

receiving an indication of a first set of words from said plurality of sets of words, one of said words within said first set of words being a homonym; and

performing a search based on said first set of words.

3. The method of claim 2, said search being performed using a plurality of words in place of said homonym.

4. The method of claim 1, said normalizing language model comprising rules for symbol representations.

5. The method of claim 1, said normalizing language model comprising rules for homonyms comprising multiple words.

6. The method of claim 1, said homonyms comprising trade names.

7. The method of claim 1, said processing being performed by a mobile device.

8. The method of claim 1, said spoken search request being an audio file received over a network.

9. The method of claim 8, said audio file being created by a mobile device.

10. The method of claim 9, at least a portion of said plurality of sets of words being transmitted to said mobile device.

11. A system comprising:

an audio input system configured to receive a spoken search request;

a language processing system comprising a text normalization language model comprising rules defining a plurality of display forms for at least one homonym;

said language processing system configured to determine a normalized set of words, at least one of said words comprising one form of said homonym; and

a presentation engine configured to prepare said a display form of said normalized set of words on a visual display, said normalized set of words comprising a display form for said homonym.

12. The system of claim 11 further comprising:

a search engine configured to perform a search using said display form of said normalized set of words.

13. The system of claim 12,

said language processing system configured to determine a plurality of normalized sets of words;

and

said presentation engine configured to prepare said plurality of normalized sets of words for visual display.

14. The system of claim 13, said search engine being configured to perform said search using a selected one of said plurality of normalized sets of words.

15. The system of claim 11, said system being embodied on a mobile device.

16. The system of claim 11, said system being embodied on a network connected server.

17. The system of claim 16, said spoken search request being generated by a mobile device.

18. The system of claim 17, said visual display being a display on said mobile device.

19. A method comprising:

receiving a spoken search request;

processing said spoken search request into a set of words using a normalizing language model comprised of word units, a portion of said word units having a homonym;

for each of said homonym, said language model having a separate word unit, said word unit being a display form for said homonym; and

generating a display form for said set of words.

20. The method of claim 19, said homonym being at least one of a group composed of:

trade names, phrases, and symbols.