Identification using audio signatures and additional characteristics

• US 9,147,399 B1
• Filed: 08/31/2012
• Issued: 09/29/2015
• Est. Priority Date: 08/31/2012
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Hart, Gregory M., Lindsay, Allan Timothy, Barton, William F., Thimsen, John Daniel
• Primary Examiner(s): Jackson, Jakieda
• Application Number:
  US13/601,551
• Time in Patent Office: 1,124 Days
• Field of Search: 704/246, 704/239, 704/275
• US Class Current: 1/1
• CPC Class Codes: G10L 17/08 Use of distortion metrics or a particular distance between probe pattern and reference templates
• CPC Class Codes: G10L 17/22 Interactive procedures; Man-machine interfaces

Abstract

Techniques for identifying users that issue audio commands based on signatures associated with the commands and additional characteristics associated with the commands. For instance, a device that includes a microphone may capture audio uttered by a user. The device, or another device, may then compare a signature associated with a generated audio signal to audio signatures associated with known users. For instance, the device may have access to multiple audio signatures, each of which is unique to a respective user that has previously interacted with the device or with another device. The device may then use this comparison to help identify the user that uttered the audio. In addition, however, the device may utilize a characteristic other than the audio signature. Using both the comparison of the audio signature to the previously received signatures along with the additional characteristic(s), the device may make a presumed identification of the user.

Claims

1. An apparatus comprising:

a microphone;

a processor; and

computer-readable media storing computer-executable instructions that, when executed on the processor, cause the processor to perform acts comprising;

generating, by the microphone, an audio signal based on audio uttered by a user and captured by the microphone;

comparing a signature of the audio signal to a previously stored signature that is associated with a particular user;

calculating a similarity between the signature of the audio signal and the previously stored signature based at least in part on the comparing;

at least partly in response to the similarity being greater than a threshold, identifying the user that uttered the audio as the particular user; and

at least partly in response to the similarity being less than the threshold;

identifying a command associated with the audio signal;

determining a characteristic associated with the command; and

attempting to identify the user that uttered the audio based at least in part on the signature of the audio signal and on the determined characteristic.

2. An apparatus as recited in claim 1, wherein the comparing comprises comparing the signature of the audio signal to multiple previously stored signatures, each of the multiple previously stored signatures being associated with a different respective user.

3. An apparatus as recited in claim 1, wherein the attempting to identify the user that uttered the audio is further based at least in part on at least one of a sequence of words within the audio uttered by the user, content to which the user has access, a schedule associated with the user, information about services with which the user has registered with, or a time of day at which the user uttered the audio.

4. An apparatus as recited in claim 1, wherein the attempting to identify the user that uttered the audio is further based at least in part on information about a room in which the apparatus resides.

5. An apparatus as recited in claim 1, wherein the attempting to identify the user that uttered the audio is further based at least in part on information about a location within a room at which the user uttered the audio.

6. An apparatus as recited in claim 1, wherein the attempting to identify the user that uttered the audio is further based at least in part on at least one of previous interactions between the user and the apparatus, between the user and an application associated with the apparatus, or between the user and a service accessible to the apparatus.

7. An apparatus as recited in claim 1, the acts further comprising receiving an indication of a location of a device associated with the user, and wherein the attempting to identify the user that uttered the audio is further based at least in part on the indication of the location of the device associated with the user.

8. An apparatus as recited in claim 1, wherein the microphone further captures additional audio from an environment, and the attempting to identify the user that uttered the audio is further based at least in part on the additional audio from the environment.

9. An apparatus as recited in claim 1, further comprising a speaker, and the acts further comprising outputting an indication of a presumed identity of the user via the speaker.

10. A method comprising:

under control of one or more computing systems configured with executable instructions,

receiving an audio signal generated by a microphone based on audio uttered by a user;

comparing a signature of the audio signal to a previously stored audio signature that is associated with a particular user;

calculating a similarity between the signature of the audio signal and the previously stored audio signature;

at least partly in response to the similarity being greater than a threshold, identifying the user that uttered the audio as the particular user; and

at least partly in response to the similarity being less than the threshold;

identifying a command associated with the audio signal;

determining a characteristic associated with the command; and

attempting to identify the user that uttered the audio based at least in part on the signature of the audio signal and on the determined characteristic.

11. A method as recited in claim 10, wherein the signature of the audio signal is based at least in part on at least one of a volume, a pitch, a tone or a frequency of the audio signal.

12. A method as recited in claim 10, wherein the receiving and the attempting occur at a device that includes the microphone that captures the audio uttered by the user and generates the audio signal.

13. A method as recited in claim 10, wherein the receiving and the attempting occur at a computing device that is remote from a device that captured the audio uttered by the user and that generated the audio signal.

14. A method as recited in claim 10, wherein the attempting comprises comparing the signature of the audio signal to multiple audio signatures associated with respectively different users, the comparing comprising comparing a volume, a pitch, a tone or a frequency of the audio signal to respective volumes, pitches, tones or frequencies of the multiple audio signatures associated with the respectively different users.

15. A method as recited in claim 10, wherein the attempting to identify the user that uttered the audio is further based at least in part on a grammar of the audio uttered by the user.

16. A method as recited in claim 10, wherein the attempting to identify the user that uttered the audio further comprises:

identifying content referenced in the audio uttered by the user; and

determining which one or more users of multiple users have access to the referenced content.

17. A method as recited in claim 10, wherein the attempting to identify the user that uttered the audio is further based at least in part on at least one of a time of day or a day on which the audio is uttered by the user.

18. A method as recited in claim 10, wherein the attempting to identify the user that uttered the audio is further based at least in part on information about a location within a room at which the audio is uttered by the user.

19. A method as recited in claim 10, wherein the audio is received by a device, and the attempting to identify the user that uttered the audio is further based at least in part on previous commands received by the device.

20. A method as recited in claim 10, wherein the audio is received by a device, and the attempting to identify the user that uttered the audio is further based at least in part on previous interactions between the device and one or more different users, between an application associated with the device and one or more different users, or between a service accessible to the device and one or more different users.

21. A method as recited in claim 10, further comprising receiving an indication of a geographical location of a device associated with the user, and wherein the attempting to identify the user that uttered the audio is further based at least in part on the indication of the geographical location of the device associated with the user.

22. A method as recited in claim 10, further comprising:

making an identification of the user based at least in part on the attempting; and

outputting an indication of the identification.

23. Non-transitory computer-readable media storing computer-executable instructions that, when executed on a processor, cause the processor to perform acts comprising:

receiving a generated audio signal that has been generated by a microphone based on audio captured by the microphone;

comparing a signature of the generated audio signal to a previously stored audio signature associated with a particular user;

calculating a similarity between the signature of the generated audio signal and the previously stored audio signature;

at least partly in response to the similarity being greater than a threshold, identifying a user that uttered audio associated with the generated audio signal as the particular user; and

at least partly in response to the similarity being less than the threshold;

identifying a command associated with the generated audio signal;

determining a characteristic associated with the command; and

attempting to identify the user that uttered the audio associated with the audio signal based at least in part on the signature of the generated audio signal and on the determined characteristic.

24. Computer-readable media as recited in claim 23, wherein the command requests a particular piece of content, and the characteristic comprises the particular piece of content.

25. Computer-readable media as recited in claim 23, wherein the command comprises a request to output an audio file, and the characteristic comprises the audio file associated with the request.

26. Computer-readable media as recited in claim 23, wherein the characteristic comprises a grammar associated with the command.

27. Computer-readable media as recited in claim 23, the acts further comprising:

making a presumed identification of the user that uttered the audio based at least in part on the attempting; and

outputting the presumed identification via a speaker.

28. Computer-readable media as recited in claim 23, the acts further comprising:

making a presumed identification of the user that uttered the audio based at least in part on the attempting; and

outputting a request to the user to confirm or deny the presumed identification via a speaker.