Speech Recognition, and Related Systems

• US 20080086311A1
• Filed: 04/06/2007
• Published: 04/10/2008
• Est. Priority Date: 04/11/2006
• Status: Abandoned Application

• Current Assignee: Digimarc Corporation
• Original Assignee: Digimarc Corporation
• Inventors: Conwell, William, Meyer, Joel
• Application Number:
  US11/697,610
• Publication Number:
  US 20080086311A1
• Time in Patent Office: Days
• US Class Current: 704/500
• CPC Class Codes: G10L 15/07 to the speaker
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 2015/227 of the speaker; Human-factor methodology
• CPC Class Codes: H04M 2250/74 with voice recognition means

Abstract

In one arrangement, information useful in understanding the content of user speech (e.g., phonemes identified by a speech recognition algorithm, data indicating the gender of the speaker, etc.) is determined at an apparatus (e.g., a cell phone), and accompanies speech data sent from that apparatus. (Steganographic encoding of the speech data can be employed to convey this information.) A receiving device can use this accompanying information to better understand the content of the speech. A great variety of other features and arrangements—some dealing with imagery rather than audio—are also detailed.

Claims

1. A method comprising the acts:

receiving audio corresponding to a user'"'"'s speech;

obtaining speech recognition data associated with said speech;

generating digital speech data corresponding to said received audio; and

transmitting the digital speech data accompanied by the speech recognition data.

2. The method of claim 1 wherein said obtaining comprises applying a speech recognition algorithm to said received audio.

3. The method of claim 2 in which the speech recognition algorithm employs recognition parameters tailored to the user.

4. The method of claim 1 wherein said obtaining comprises obtaining data indicating a language of said speech.

5. The method of claim 1 wherein said obtaining comprises obtaining data indicating a gender of said user.

6. The method of claim 1 wherein said transmitting includes steganographically encoding said digital speech data with said speech recognition data.

7. The method of claim 1, performed by a wireless communications device.

8. The method of claim 1 wherein said transmitting further includes transmitting context information with said digital speech data and said speech recognition data.

9. A method performed at a first location, using a speech signal provided from a remote location, comprising the acts:

obtaining speech recognition data conveyed with the speech signal; and

applying a speech recognition algorithm to said speech signal, employing the speech recognition data conveyed therewith.

10. The method of claim 9, wherein said obtaining comprises decoding speech recognition data steganographically encoded in said speech signal.

11. The method of claim 9 that further includes, at the remote location and prior to the provision of said speech signal to said first location, applying a preliminary speech recognition algorithm to said speech signal, and conveying speech recognition data resulting therefrom with said speech signal.

12. The method of claim 11 in which said conveying comprises steganographically encoding said speech recognition data into said speech signal.

13. The method of claim 11 in which said preliminary speech recognition algorithm employs a model especially tailored to a speaker of said speech.

14. The method of claim 9 that further comprises transmitting to a web service a result of said speech recognition algorithm, together with context information.

15. The method of claim 14 that further includes receiving at a user device certain information responsive to said transmission to the web service, and dependent on said context information.

16. In a telecommunications method that includes sensing speech from a speaker, and relaying speech data corresponding thereto to a remote location, an improvement comprising conveying auxiliary information with said speech data, said auxiliary information comprising at least one of the following:

data indicating a language of said speech, data indicating an age of said speaker, or data indicating a gender of said speaker.

17. The method of claim 16 in which said conveying comprises steganographically encoding said speech data to convey said auxiliary information.

18. A method comprising:

at a first, battery-powered, wireless device, performing an initial recognition operation on received audio or image content;

conveying a representation of said content, together with data resulting from said initial recognition operation, from said first device to a second, remotely located, device; and

at said second device, performing a further recognition operation on said representation of content, said further operation making use of data resulting from said initial operation.

19. The method of claim 18, performed on image content.

20. A mobile handset including a microphone and a speech recognition system, characterized in that a processor thereof changes the handset between different modes of operation depending on assessment of speech recognition accuracy.

21. A method using a handheld wireless communications device that includes a camera system which captures raw image data, converts same to RGB data, and compresses the RGB data, the method further including performing at least a partial fingerprint determination operation on the raw image data prior to said conversion-to-RGB and prior to said compression, and sending resultant fingerprint information from said device to a remote system.

22. The method of claim 21 that further comprises performing a further fingerprint determination operation on the sent information at said remote system.

23. The method of claim 21 that further comprises capturing plural frames of image information using said sensor, and combining raw image data from said frames to yield higher quality data prior to performing said fingerprint determination operation on the raw image-data.

24. A method of fingerprint determination comprising:

at a wireless communications device, capturing audio;

performing a partial fingerprint determination on data corresponding to said captured audio;

transmitting results from said partial fingerprint determination to a remote system; and

performing a further fingerprint determination on said remote system.

25. A method comprising:

capturing an image including a face using a camera system of a handheld wireless communications device;

performing a partial signature calculation characterizing the face in said image, using a processor in said handheld wireless communications device;

transmitting data resulting from said partial signature calculation to a remote system;

performing a further signature calculation on the remote system; and

using resultant signature data to seek a match between said face and a reference database of facial image data.