AUTOMATED VERBAL FLUENCY ASSESSMENT

• US 20150058013A1
• Filed: 03/14/2013
• Published: 02/26/2015
• Est. Priority Date: 03/15/2012
• Status: Active Grant

• Current Assignee: Regents of The University of Minnesota (University of Minnesota)
• Original Assignee: Regents of The University of Minnesota (University of Minnesota)
• Inventors: Pakhomov, Serguei V.S., Hemmy, Laura Sue, Lim, Kelvin O.
• Granted Patent: US 9,576,593 B2
• Time in Patent Office: Days
• US Class Current: 704/243
• CPC Class Codes: G10L 15/02 Feature extraction for speech recognition; Selection of recognition unit
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 2025/783 based on threshold decision
• CPC Class Codes: G10L 25/48 specially adapted for particular use
• CPC Class Codes: G10L 25/66 for extracting parameters related to health condition detecting or measuring for diagnostic purposes A61B5/00
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10
• CPC Class Codes: G10L 25/87 Detection of discrete points within a voice signal

Abstract

Techniques are described for calculating one or more verbal fluency scores for a person. An example method includes classifying, by a computing device, samples of audio data of speech of a person, based on amplitudes of the samples, into a first class of samples including speech or sound and a second class of samples including silence. The method further includes analyzing the first class of samples to determine a number of words spoken by the person, and calculating a verbal fluency score for the person based at least in part on the determined number of words spoken by the person.

Claims

1. A method comprising:

classifying, by a computing device, samples of audio data of speech of a person, based on amplitudes of the samples and on a silence threshold, into a first class of samples including speech or sound and a second class of samples including silence;

analyzing the first class of samples to determine a number of words spoken by the person; and

calculating a verbal fluency score for the person based at least in part on the determined number of words spoken by the person.

2. The method of claim 1, wherein classifying the samples comprises:

sorting the samples of the audio data in an order defined by the amplitudes of the samples of the audio data;

determining the silence threshold based on the amplitudes of the samples of the audio data;

classifying samples having amplitudes above the silence threshold as belonging to the first class; and

classifying samples having amplitudes below the silence threshold as belonging to the second class.

3. The method of claim 2, wherein determining the silence threshold comprises:

calculating linear regressions of the sorted samples in the sorted order; and

determining the silence threshold when a calculated linear regression has a slope that exceeds a predetermined value.

4. The method of claim 3, wherein the predetermined value comprises −

0.2.

5. The method of claim 1, wherein analyzing the first class of samples comprises determining a first subset of samples of the first class including speech sound and a second subset of samples of the first class including non-speech sound.

6. The method of claim 5, further comprising determining the number of words as a number of contiguous samples in the audio data belonging to the first subset that start with a sample above the silence threshold and end with a sample below the silence threshold.

7. The method of claim 5, wherein determining the first subset and the second subset comprises:

classifying contiguous samples in the first class of the audio data for which a fundamental frequency can be calculated as belonging to the first subset; and

classifying contiguous samples in the first class of the audio data for which a fundamental frequency cannot be calculated as belonging to the second subset.

8. The method of claim 1, further comprising:

determining a number of pauses as a number of contiguous samples in the second class that start with a sample below the silence threshold and end with a sample below the silence threshold.

9. The method of claim 8, further comprising:

measuring a duration associated with each pause of the pauses;

calculating an average duration comprising a mean value of the measured durations; and

calculating a standard deviation of the measured durations from the average duration.

10. The method of claim 1, further comprising:

classifying, by the computing device, second samples of second audio data of speech of the person, based on second amplitudes of the second samples and on a second silence threshold, into the first class and the second class;

calculating a second verbal fluency score based at least in part on the number of words spoken by the person; and

calculating a learning score based at least in part on a change from the verbal fluency score to the second verbal fluency score.

11. The method of claim 10, wherein the silence threshold and the second silence threshold comprise equal values.

12. The method of claim 10, wherein calculating the learning score further comprises:

plotting at least the verbal fluency score and the second verbal fluency score on a graph; and

calculating a slope associated with the graph.

13. The method of claim 1, further comprising:

receiving the samples as at least a portion of a verbal fluency test of the person.

14. The method of claim 13, further comprising:

outputting the verbal fluency score.

15. A device comprising:

a memory storing instructions; and

one or more processors configured to execute the instructions, wherein execution of the instructions causes the one or more processors to;

classify samples of audio data of speech of a person, based on amplitudes of the samples and on a silence threshold, into a first class of samples including speech or sound and a second class of samples including silence;

analyze the first class of samples to determine a number of words spoken by the person; and

calculate a verbal fluency score for the person based at least in part on the determined number of words spoken by the person.

16. The device of claim 15, wherein to classify the samples, the one or more processors are configured to:

sort the samples of the audio data in an order defined by the amplitudes of the samples of the audio data;

determine the silence threshold based on the amplitudes of the samples of the audio data;

classify samples having amplitudes above the silence threshold as belonging to the first class; and

classify samples having amplitudes below the silence threshold as belonging to the second class.

17. The device of claim 15, wherein to analyze the first class of samples, the one or more processors are configured to:

determine a first subset of samples of the first class including speech sound and a second subset of samples of the first class including non-speech sound.

18. The device of claim 17, wherein to determine the first subset and the second subset, the one or more processors are configured to:

classify contiguous samples in the first class of the audio data for which a fundamental frequency can be calculated as belonging to the first subset; and

classify contiguous samples in the first class of the audio data for which a fundamental frequency cannot be calculated as belonging to the second subset.

19. The device of claim 15, wherein the one or more processors are further configured to:

classify second samples of second audio data of speech of the person, based on second amplitudes of the second samples and on a second silence threshold, into the first class and the second class;

calculate a second verbal fluency score based at least in part on the number of words spoken by the person; and

calculate a learning score based at least in part on a change from the verbal fluency score to the second verbal fluency score.

20. A computer-readable medium comprising instructions that, when executed, cause a processor to:

classify samples of audio data of speech of a person, based on amplitudes of the samples and on a silence threshold, into a first class of samples including speech or sound and a second class of samples including silence;

analyze the first class of samples to determine a number of words spoken by the person; and

calculate a verbal fluency score for the person based at least in part on the determined number of words spoken by the person.