Method for training an automatic speech recognition system

• US 10,049,658 B2
• Filed: 03/07/2013
• Issued: 08/14/2018
• Est. Priority Date: 03/07/2013
• Status: Expired due to Fees

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Pinto, Joel, Anastasiadis, Josef Damianus, Willett, Daniel
• Primary Examiner(s): Leland, III, Edwin S
• Application Number:
  US14/773,142
• Publication Number:
  US 20160027435A1
• Time in Patent Office: 1,986 Days
• Field of Search: 704244
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/07 to the speaker
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/0635 updating or merging of old and new templates; Mean values; Weighting
• CPC Class Codes: G10L 2015/0638 Interactive procedures

Abstract

A system and method for speech recognition is provided. Embodiments may include receiving, at a first computing device, a far-talk signal from a far-talk computing device, the far-talk signal transmitted using a first channel and corresponding to an audible sound. Embodiments may further include receiving, at the first computing device, a near-talk signal from a near-talk computing device, the near-talk signal transmitted using a second channel and corresponding to the audible sound, wherein the far-talk signal and the near-talk signal are received during an enrollment phase of a far-talk speech recognition system. Embodiments may also include updating, at the first computing device, one or more models associated with a far-talk speech recognition system based upon, at least in part, one or more characteristics of the far-talk signal and one or more characteristics of the near-talk signal.

Claims

1. A computer-implemented method comprising:

receiving, at a first server computing device via a network, a far-talk signal from a far-talk computing device having at least one processor, the far-talk signal transmitted using a first channel and corresponding to an audible sound of a user, wherein the network is at least one of the Internet or a local area network;

receiving, at the first server computing device via the network, a near-talk signal from a near-talk computing device having at least one processor, the near-talk signal transmitted using a second channel and corresponding to the audible sound of the user, wherein the far-talk signal and the near-talk signal are received during an enrollment phase of a far-talk speech recognition system; and

updating, at the first server computing device, one or more voice models associated with a far-talk speech recognition system that best match a speech pattern of the user with the one or more voice models utilizing, at least in part, one or more characteristics of the far-talk signal and one or more characteristics of the near-talk signal.

2. The method of claim 1, wherein updating includes estimating one or more references for applying unsupervised adaptation techniques associated with the far-talk speech recognition system.

3. The method of claim 1, wherein updating includes determining a non linear mapping between the far-talk signal and the near-talk signal.

4. The method of claim 1, wherein updating includes determining at least one room impulse response characteristic.

5. The method of claim 1, wherein updating includes optimizing at least one beam-forming setting.

6. The method of claim 1 wherein the near-talk device is a mobile phone.

7. The method of claim 6, wherein the far-talk device is an in-car device.

8. The method of claim 1, wherein receiving a far-talk signal includes receiving an identification code associated with the far-talk computing device.

9. The method of claim 1, wherein receiving a near-talk signal includes receiving an identification code associated with the near-talk computing device.

10. The method of claim 1, further comprising:

transmitting via the network, and using the first server computing device, the one or more models to the far-talk computing device.

11. The method of claim 1 wherein the near-talk device is a remote control and the far-talk device is a television or a device associated with a television.

12. A system comprising:

a first server computing device having at least one processor configured to receive, via a network, a far-talk signal from a far-talk computing device having at least one processor, the far-talk signal transmitted using a first channel and corresponding to an audible sound of a user, wherein the network is at least one of the Internet or a local area network, the first server computing device having the at least one processor being further configured to receive a near-talk signal from a near-talk computing device having at least one processor, the near-talk signal transmitted using a second channel and corresponding to the audible sound of the user, wherein the far-talk signal and the near-talk signal are received during an enrollment phase of a far-talk speech recognition system, the first server computing device having the at least one processor being further configured to update one or more voice models associated with a far-talk speech recognition system that best match a speech pattern of the user with the one or more voice models utilizing, at least in part, one or more characteristics of the far-talk signal and one or more characteristics of the near-talk signal.

13. The system of claim 12, wherein updating includes estimating one or more references for applying unsupervised adaptation techniques associated with the far-talk speech recognition system.

14. The system of claim 12, wherein updating includes determining a non linear mapping between the far-talk signal and the near-talk signal.

15. The system of claim 12, wherein updating includes determining at least one room impulse response characteristic.

16. The system of claim 12, wherein updating includes optimizing at least one beam-forming setting.

17. The system of claim 12, wherein the near-talk device is at least one of a mobile phone and a remote control device.

18. The system of claim 12, wherein the far-talk device is at least one of a television and an in-car device.

19. The system of claim 12, wherein receiving a far-talk signal includes receiving an identification code associated with the far-talk computing device.

20. The system of claim 12, wherein receiving a near-talk signal includes receiving an identification code associated with the near-talk computing device.

21. A computer-implemented method comprising:

transmitting, via a network, a far-talk signal from a far-talk computing device having at least one processor to a server computing device, the far-talk signal transmitted using a first channel and corresponding to an audible sound of a user, wherein the network is at least one of the Internet or a local area network;

transmitting, via the network, a near-talk signal from a near-talk computing device having at least one processor to a server computing device, the near-talk signal transmitted using a second channel and corresponding to the audible sound of the user, wherein the far-talk signal and the near-talk signal are received during an enrollment phase of a far-talk speech recognition system; and

updating, by the server computing device, one or more voice models associated with a far-talk speech recognition system that best match a speech pattern of the user with the one or more voice models utilizing, at least in part, one or more characteristics of the far-talk signal and one or more characteristics of the near-talk signal.