Method of classifying respiratory sounds

• US 5,879,313 A
• Filed: 01/21/1997
• Issued: 03/09/1999
• Est. Priority Date: 12/22/1994
• Status: Expired due to Fees

• Current Assignee: SNAP Diagnostics, LLC
• Original Assignee: SNAP Laboratories, LLC
• Inventors: Raviv, Gil, Weingarten, Charles Z.
• Primary Examiner(s): Hindenburg, Max
• Assistant Examiner(s): WINGOOD, PAMELA LYNN
• Application Number:
  US08/786,631
• Time in Patent Office: 777 Days
• Field of Search: 600/587, 600/552, 600/586, 600/595
• US Class Current: 600/595
• CPC Class Codes: A61B 7/003 Detecting lung or respiration noise

Abstract

The invention provides a method of classifying respiratory sounds. The method provides the steps of selecting a first set of respiratory sounds and manually determining a classification content of the first set of respiratory events. The method further includes extrapolating the classification content of the first set of respiratory events to an at least second set of respiratory events.

Claims

1. A method of analyzing respiratory sounds comprising the steps of:

selecting a first set of respiratory events;

manually determining a classification content of the first set of respiratory events; and

extrapolating the classification content of the first set of respiratory events to an at least second set of respiratory events.

2. The method as in claim 1 further comprising the step of classifying at least some of the respiratory sounds of the first set of respiratory events as snoring events.

3. The method as in claim 1 further comprising the step of classifying at least some of the respiratory sounds of the first set of respiratory events as apnea.

4. The method as in claim 1 further comprising the step of fast fourier transforming at least a portion of the respiratory events of the first set of respiratory events.

5. The method as in claim 1 wherein the step of manually determining a classification content further comprises the step of visually scanning a graphical representation of the fast fourier transform of the breathing events for frequency patterns such as harmonics or diffuse patterns.

6. The method as in claim 5 further comprising the step of listening to respiratory events of the first respiratory event set and classifying the respiratory event based upon harmonic frequency patterns and the audio sound of the respiratory event.

7. The method as in claim 1 further comprising the step of randomly selecting the first set of respiratory events.

8. The method of claim 2 further comprises determining an amplitude distribution index as a ratio of two selected snoring amplitude values derived from selected ones of the respiratory events classified as snoring events.

9. The method of claim 2 wherein snoring events are ordered according to descending FFT amplitude of each snoring event.

10. The method of claim 8 further comprising grouping the ordered snoring events and determining an amplitude distribution index as a ratio of amplitudes of two selection groups.

11. A method of analyzing respiratory sounds comprising the steps of:

selecting a first set of respiratory events to provide pre-classified respiratory events;

selecting a subset of the pre-classified respiratory events;

manually classifying the events of the subset;

comparing the manually classified events and pre-classified events; and

tabulating the results of the comparison in a table.

12. The method as in claim 11 further comprising the step of correcting the full set of pre-classified respiratory events based upon location within the table.

13. The method as in claim 12 further comprising the step of defining the set of pre-classified respiratory events as automatically classified respiratory events.

14. The method as in claim 12 further comprising the step of defining the set of pre-classified respiratory events as manually classified respiratory events.

15. The method as in claim 12 further comprising the step of defining the set of pre-classified respiratory events as either a set of manually or automatically classified respiratory events.

16. The method as in claim 12 further comprising the step of defining at least some of the pre-classified respiratory events as snoring.

17. The method as in claim 12 further comprising the step of defining at least some of the pre-classified respiratory events as apnea.

18. A method of analyzing respiratory sounds in a set of respiratory sounds comprising the steps of:

selecting at least one respiratory event from the respiratory sounds;

measuring the loudness of the at least one event;

calculating an average relative snoring loudness by averaging the loudness of a set of pre-classified snores; and

comparing the measured loudness to the average relative snoring loudness to determine the relative snoring loudness of the respiratory event.

19. The method as in claim 18 further comprising the step of calculating a maximum or average or weighted FFT amplitude over a predefined range of frequencies for loudness measurement.

20. The method as in claim 18 further comprising the step of calculating a maximum relative loudness on a pre-selected subset of snores by averaging a loudest few percent of the snores, and calculating the relative loudness of the averaged snores.

21. The method as in claim 18 further comprising the step of calculating a maximum relative loudness on a pre-selected subset of snores by selecting a snore having a maximum loudness within the pre-selected subset of snores and calculating the relative loudness of the loudest snore.

22. The method as in claim 18 further comprising the step of calculating the relative loudness in decibels by the expression "R.L=20 log 10 (S/NA)" or "R.L.=10 log 10 (S/NA)", where S is the max FFT amplitude of the snore and NA is the average or maximum loudness of the normal breathing in the reference subset.

23. A method of analyzing respiratory sounds comprising the steps of:

selecting a first group of respiratory events classified as snores;

determining a maximum or average first loudness value for the first group of snores;

determining a maximum or average second loudness value for a second group of snores; and

calculating a loudness difference or ratio between the first and second loudness values.

24. The method as in claim 23 further comprising the step of defining the first group of snores as velum-like snores and the second group of snores as non-velum snores.

25. The method as in claim 23 further comprising the step of calculating a loudness threshold where the percentage of velum-like snores with loudness above the loudness threshold exceeds a predetermined percentage threshold and where the percentage of velum-like snores with loudness below the loudness threshold is less than the predetermined percentage threshold.

26. The method as in claim 23 further comprising the step of calculating an amplitude distribution index by calculating a ratio between a maximum relative loudness and the loudness threshold.

27. A method of analyzing respiratory sounds comprising the steps of:

selecting of group of respiratory events;

fast fourier transforming each respiratory event of the group of respiratory events to produce a set of FFT values for each respiratory event;

identifying a relative maxima among the FFT values of each snoring event within the group of respiratory events;

identifying a relative maxima FFT value among the normal respiratory events with the group of respiratory events; and

calculating a logarithmic value of a ratio determined by dividing the relative maxima FFT value among snoring events by an average of maxima FFT values of the normal respiratory events.

28. The method as in claim 27 further comprising the step of multiplying the logarithmic value by the integer value of 10 or 20.