Impairment detection using speech

• US 8,478,596 B2
• Filed: 11/28/2005
• Issued: 07/02/2013
• Est. Priority Date: 11/28/2005
• Status: Active Grant

• Current Assignee: Rakuten Group, Inc.
• Original Assignee: Verizon Business Global LLC (Verizon Communications Inc.)
• Inventors: Schultz, Paul T.
• Primary Examiner(s): Armstrong, Angela A
• Application Number:
  US11/287,358
• Publication Number:
  US 20070124135A1
• Time in Patent Office: 2,773 Days
• Field of Search: 704/201, 704/270, 704/275, 704/273, 704/200
• US Class Current: 704/270
• CPC Class Codes: G10L 17/26 Recognition of special voice characteristics, e.g. for use in lie detectors; Recognition of animal voices

Abstract

A device may include logic configured to receive a first speech input from a party, to compare the first speech input to a second speech input to produce a result, and to determine if the party is impaired based on the result.

Claims

1. A device, comprising:

a memory to store a plurality of speech inputs,

each speech input, of the stored plurality of speech inputs, being received from a corresponding party of a plurality of parties; and

a processor to;

receive a first speech input from a particular party of the plurality of parties,

access the speech input, of the stored plurality of speech inputs, received from the particular party based on receiving the first speech input,

the accessed speech input including a particular phrase,

provide a prompt to the particular party for the particular phrase,

receive a second speech input from the particular party based on providing the prompt to the particular party,

determine speech characteristics of the particular phrase included in the accessed speech input and speech characteristics of the received second speech input,

the determined speech characteristics of the particular phrase including a number of pauses in the particular phrase, a frequency variability of a portion of the particular phrase, and at least one of;

a non-fluency of a portion of the particular phrase, a speaking rate associated with the particular phrase, a length of a pause in the particular phrase, a misarticulation of a portion of the particular phrase, or a vocal intensity of a portion of the particular phrase, and

the determined speech characteristics of the received second speech input including a number of pauses in the received second speech input, a frequency variability of a portion of the received second speech input, and at least one of;

a non-fluency of a portion of the received second speech input, a speaking rate associated with the received second speech input, a length of a pause in the received second speech input, a misarticulation of a portion of the received second speech input, or a vocal intensity of a portion of the received second speech input,

compare the determined speech characteristics of the particular phrase to the determined speech characteristics of the received second speech input to produce a result, and

determine whether the party is impaired based on the result.

2. The device of claim 1, where, when comparing the determined speech characteristics of the particular phrase to the determined speech characteristics of the received second speech input, the processor is to:

evaluate, using biometric voice verification techniques, the determined speech characteristics of the particular phrase and the determined speech characteristics of the received second speech input.

3. The device of claim 1, where, when determining the speech characteristics of the particular phrase and the speech characteristics of the received second speech input, the processor is to:

interpret speech segments comprising the accessed speech input and the received second speech input.

4. The device of claim 1, where the processor is further to:

receive the accessed speech input when the particular party is not impaired.

5. The device of claim 1, where the result represents a score indicative of a deviation between a segment of the received second speech input and a corresponding segment of the accessed speech input, and

where the segment of the accessed speech input includes the particular phrase.

6. The device of claim 1, where the processor is further to:

receive the accessed speech input via a first telephone call placed by the particular party, and

receive the second speech input via a second telephone call placed by the particular party.

7. The device of claim 1, where the processor is further to:

identify a schedule for receiving speech inputs from the particular party, and

determine whether the second speech input is received according to the schedule.

8. A method, comprising:

receiving, by a device, speech inputs from a plurality of parties;

storing, by the device, the received speech inputs;

receiving, by the device, a first speech input from a particular party of the plurality of parties;

accessing, by the device, a particular speech input, from the stored speech inputs, received from the particular party, based on the first speech input,

the particular speech input including a particular phrase;

providing, by the device, a prompt to the particular party for the particular phrase;

receiving, by the device, a second speech input from the particular party based on providing the prompt;

determining, by the device, speech characteristics of the particular phrase and speech characteristics of the second speech input,

the speech characteristics of the particular phrase including a number of pauses in the particular phrase and two or more of;

a vocal intensity of a portion of the particular phrase, a speaking rate associated with the particular phrase, a length of a pause in the particular phrase, a misarticulation of a portion of the particular phrase, a non-fluency of a portion of the particular phrase, or a frequency variability of a portion of the particular phrase, and

the speech characteristics of the second speech input including a number of pauses in the second speech input and two or more of;

a vocal intensity of a portion of the second speech input, a speaking rate associated with the second speech input, a length of a pause in the second speech input, a misarticulation of a portion of the second speech input, a non-fluency of a portion of the second speech input, or a frequency variability of a portion of the second speech input;

comparing, by the device, the determined speech characteristics of the second speech input to the determined speech characteristics of the particular phrase to produce a score; and

determining, by the device, a condition of the particular party based on the score.

9. The method of claim 8, where determining the condition comprises:

evaluating, using voice recognition logic or biometric voice verification logic, the determined speech characteristics of the second speech input and the determined speech characteristics of the particular phrase.

10. The method of claim 8, where comparing the determined speech characteristics of the second speech input to the determined speech characteristics of the particular phrase comprises:

identifying a deviation between at least one segment of the accessed particular speech input and a corresponding at least one segment of the second speech input,

the at least one segment of the accessed particular speech input including the particular phrase; and

producing the score based on the deviation.

11. The method of claim 8, where the accessed particular speech input includes a plurality of segments and the second speech input includes a corresponding plurality of segments, and

where comparing the determined speech characteristics of the second speech input to the determined speech characteristics of the particular phrase comprises:

comparing the corresponding plurality of segments to the plurality of segments to produce a plurality of results.

12. The method of claim 8, further comprising:

determining an identity of the particular party based on the first speech input prior to receiving the second speech input from the particular party.

13. The method of claim 12, where the particular party is located at a first location, and

where determining the identity is performed at a second location.

14. The method of claim 8, where determining the condition of the particular party comprises:

determining whether the particular party is impaired based on the score,

where the score represents a confidence level that the particular party is impaired.

15. The method of claim 8, where receiving the speech inputs comprises:

receiving the particular speech input from the particular party when the particular party is unimpaired.

16. A system comprising:

a memory to store;

instructions, and

store a plurality of speech inputs received from a plurality of users; and

a processor to execute instructions to;

receive a first speech input from a particular user of the plurality of users,

identify the particular user based on the first speech input,

access a particular speech input, from the stored plurality of speech inputs, based on identifying the particular user,

the particular speech input including a particular phrase, provide a prompt to the particular user for the particular phrase, receive a second speech input from the particular user based on providing the prompt,

determine characteristics of the particular phrase and characteristics of the second speech input,

the determined characteristics of the particular phrase including a number of pauses in the particular phrase, a frequency variability of a portion of the particular phrase, and at least one of;

a non-fluency of a portion of the particular phrase, a speaking rate associated with the particular phrase, a length of a pause in the particular phrase, a misarticulation of a portion of the particular phrase, or a vocal intensity of a portion of the particular phrase, and

the determined characteristics of the second speech input including a frequency variability of a portion of the second speech input, a number of pauses in the second speech input, and at least one of;

a non-fluency of a portion of the second speech input, a speaking rate associated with the second speech input, a length of a pause in the second speech input, a misarticulation of a portion of the second speech input, or a vocal intensity of a portion of the second speech input,

compare the determined characteristics of the particular speech input to the determined characteristics of the second speech input, and

determine a condition of the particular user based on comparing the determined characteristics of the particular speech input to the determined characteristics of the second speech input.

17. The system of claim 16, where, when comparing the determined characteristics of the particular speech input to the determined characteristics of the second speech input, the processor is further to execute the instructions to:

process the particular speech input and the second speech input via speech recognition logic or biometric voice verification logic.

18. The system of claim 16, where the processor is further to execute the instructions to:

produce a score based on comparing the determined characteristics of the particular speech input to the determined characteristics of the second speech input, and

determine whether the particular user is impaired based on the score.

19. The system of claim 18, where the processor is further to execute the instructions to use the score to determine whether the particular user is in compliance with a criteria.

20. The system of claim 16, where, when comparing the determined characteristics of the particular speech input to the determined characteristics of the second speech input, the processor is further to execute the instructions to compare:

one or more of;

the vocal intensity of the portion of the particular phrase, the speaking rate associated with the particular phrase, the length of the pause in the particular phrase, the misarticulation of the portion of the particular phrase, or the non-fluency of the portion of the particular phrase, to

corresponding one or more of;

the vocal intensity of the portion of the second speech input, the speaking rate associated with the second speech input, the number of pauses in the second speech input, the length of the pause in the second speech input, the misarticulation of the portion of the second speech input, or the non-fluency of the portion of second speech input.

21. A non-transitory computer-readable medium for storing instructions, the instructions comprising:

one or more instructions which, when executed by a device, cause the device to receive speech inputs from a plurality of users;

one or more instructions which, when executed by the device, cause the device to store the speech inputs;

one or more instructions which, when executed by the device, cause the device to receive a first speech input from a particular user, of the plurality of users, during a telephone call from the particular user;

one or more instructions which, when executed by the device, cause the device to identify the particular user based on the first speech input;

one or more instructions which, when executed by the device, cause the device to access a particular speech input, from the stored speech inputs, based on identifying the particular user,

the particular speech input including a particular phrase;

one or more instructions which, when executed by the device, cause the device to provide a prompt for the particular phrase;

one or more instructions which, when executed by the device, cause the device to receive a second speech input from the particular user based on providing the prompt;

one or more instructions which, when executed by the device, cause the device to compare the particular phrase to the second speech input,

the one or more instructions to compare the particular phrase to the second speech input comprising one or more instructions to compare;

a number of pauses in the particular phrase and two or more of;

a non-fluency of a portion of the particular phrase, a speaking rate associated with the particular phrase, a misarticulation of a portion of the particular phrase, a vocal intensity of a portion of the particular phrase, a frequency variability of a portion of the particular phrase, or a length of a pause in the particular phrase to

a number of pauses in the second speech input and two or more of;

a non-fluency of a portion of the second speech input, a speaking rate associated with the second speech input, a misarticulation of a portion of the second speech input, a vocal intensity of a portion of the second speech input, a frequency variability of a portion of the second speech input, or a length of a pause in the second speech input; and

one or more instructions which, when executed by the device, cause the device to determine whether the particular user is impaired based on comparing the particular phrase to the second speech input.

22. The non-transitory computer-readable medium of claim 21, the instructions further comprising:

one or more instructions to generate a score based on comparing the particular phrase to the second speech input; and

one or more instructions to provide the score to another user associated with the device.

23. The non-transitory computer-readable medium of claim 21, the one or more instructions to compare the particular phrase to the second speech input further comprising:

one or more instructions to compare the number of pauses in the particular phrase to the number of pauses in the second speech input.

24. The non-transitory computer-readable medium of claim 21, the one or more instructions to compare the particular phrase to the second speech input further comprising:

one or more instructions to compare the speaking rate in the particular phrase to the speaking rate in the second speech input.

25. The non-transitory computer-readable medium of claim 21, the one or more instructions to compare the particular phrase to the second speech input further comprising:

one or more instructions to compare the frequency variability of the portion of the particular phrase to the frequency variability of the portion of the second speech input.

26. The non-transitory computer-readable medium of claim 21, the instructions further comprising:

one or more instructions to receive each speech input, of the speech inputs, from a corresponding user, of the plurality of users, when the user is not impaired.