Estimating lung volume by speech analysis
Described embodiments include an apparatus that includes a network interface and a processor. The processor is configured to receive, via the network interface, a speech signal that represents speech uttered by a subject, the speech including one or more speech segments, divide the speech signal into multiple frames, such that one or more sequences of the frames represent the speech segments, respectively, compute respective estimated total volumes of air exhaled by the subject while the speech segments were uttered, by, for each of the sequences, computing respective estimated flow rates of air exhaled by the subject during the frames belonging to the sequence and, based on the estimated flow rates, computing a respective one of the estimated total volumes of air, and, in response to the estimated total volumes of air, generate an alert. Other embodiments are also described.
1. A system, comprising:
a server, comprising a network interface; and
one or more processors, at least one of which belongs to the server, the processors being configured to cooperatively carry out a process that includes:
receiving from an audio-receiving device, via the network interface, a speech signal that represents speech uttered by a subject,
dividing at least part of the speech signal into multiple frames representing respective portions of the speech,
applying a function, which was computed during a calibration procedure, to one or more features of each of the frames so as to compute respective estimated volumetric flow rates of air exhaled by the subject while uttering the portions of the speech,
computing, based on the estimated volumetric flow rates, an estimated total volume of air exhaled by the subject while uttering the portions of the speech, and
in response to the estimated total volume of air, communicating an alert from the server to the audio-receiving device.
2. The system according to claim 1 , wherein a duration of each of the frames is between 5 and 40 ms.
3. The system according to claim 1 ,
wherein the speech includes multiple speech segments separated from each other by respective pauses,
wherein the portions of the speech belong to one of the segments,
wherein the process further includes computing respective other estimated total volumes of air exhaled by the subject while uttering other ones of the segments, and
wherein communicating the alert includes communicating the alert in response to the estimated total volume of air and the other estimated total volumes of air.
4. The system according to claim 1 , wherein the features include a frame energy.
5. The system according to claim 1 , wherein the function includes a polynomial function of at least one of the features, an order of the polynomial function being at least two.
6. The system according to claim 3 ,
wherein the process further includes computing one or more statistics of the estimated total volume of air and the other estimated total volumes of air, and
wherein communicating the alert includes communicating the alert in response to at least one of the statistics deviating from a baseline statistic.
7. The system according to claim 6 , wherein the speech is uttered by the subject while the subject is lying down.
8. The system according to claim 7 , wherein the process further includes:
receiving another speech signal, which represents other speech uttered by the subject while the subject is not lying down, and
computing the baseline statistic from the other speech signal.
9. The system according to claim 6 , wherein the at least one of the statistics is a statistic selected from the group of statistics consisting of: a mean, a standard deviation, and a percentile.
10. The system according to claim 6 , wherein the process further includes computing the baseline statistic from another speech signal that represents prior speech of the subject.
11. The system according to claim 1 , wherein the speech is captured by an audio sensor, and wherein the process further includes, prior to computing the estimated total volume of air, normalizing the speech signal to account for a position of the audio sensor relative to a mouth of the subject, based on images of the mouth that were acquired while the speech was uttered.
12. The system according to claim 1 , wherein the audio-receiving device comprises a screen, and wherein the process further includes displaying the alert on the screen.
13. The system according to claim 1 , wherein communicating the alert includes communicating the alert by placing a call or sending a text message.
14. The system according to claim 1 , wherein the process further includes performing the calibration procedure by:
receiving a calibration speech signal representing other speech uttered by the subject,
receiving an airflow-rate signal representing measured flow rates of air exhaled by the subject while uttering the other speech, and
using the calibration speech signal and the airflow-rate signal, computing the function.
15. The system according to claim 14 , further comprising a sound-capturing unit configured to record the calibration speech signal.
16. The system according to claim 14 , further comprising a pneumotach configured to measure the airflow-rate signal.
17. The system according to claim 16 , further comprising a sound-capturing unit disposed inside of the pneumotach and configured to record the calibration speech signal.
18. A method, comprising:
receiving, from an audio-receiving device, a speech signal that represents speech uttered by a subject;
dividing at least part of the speech signal into multiple frames representing respective portions of the speech;
by applying a function, which was computed during a calibration procedure, to one or more features of each of the frames, computing respective estimated volumetric flow rates of air exhaled by the subject while uttering the portions of the speech;
based on the estimated volumetric flow rates, computing an estimated total volume of air exhaled by the subject while uttering the portions of the speech; and
in response to the estimated total volume of air, communicating an alert from a server to the audio-receiving device.
19. The method according to claim 18 , wherein a duration of each of the frames is between 5 and 40 ms.
20. The method according to claim 18 ,
wherein the speech includes multiple speech segments separated from each other by respective pauses,
wherein the portions of the speech belong to one of the segments,
wherein the method further comprises computing respective other estimated total volumes of air exhaled by the subject while uttering other ones of the segments, and
wherein communicating the alert comprises communicating the alert in response to the estimated total volume of air and the other estimated total volumes of air.
21. The method according to claim 18 , wherein the features include a frame energy.
22. The method according to claim 18 , wherein the function includes a polynomial function of at least one of the features, an order of the polynomial function being at least two.
23. The method according to claim 20 ,
wherein the method further comprises computing one or more statistics of the estimated total volume of air and the other estimated total volumes of air, and
wherein communicating the alert comprises communicating the alert in response to at least one of the statistics deviating from a baseline statistic.
24. The method according to claim 23 , wherein the speech is uttered by the subject while the subject is lying down.
25. The method according to claim 24 , further comprising:
receiving another speech signal, which represents other speech uttered by the subject while the subject is not lying down; and
computing the baseline statistic from the other speech signal.
26. The method according to claim 23 , wherein the at least one of the statistics is a statistic selected from the group of statistics consisting of: a mean, a standard deviation, and a percentile.
27. The method according to claim 23 , further comprising computing the baseline statistic from another speech signal that represents prior speech of the subject.
28. The method according to claim 18 , wherein the speech is captured by an audio sensor, and wherein the method further comprises, prior to computing the estimated total volume of air, normalizing the speech signal to account for a position of the audio sensor relative to a mouth of the subject, based on images of the mouth that were acquired while the speech was uttered.
29. The method according to claim 18 , wherein the audio-receiving device includes a screen, and wherein the method further comprises displaying the alert on the screen.
30. The method according to claim 18 , wherein communicating the alert includes communicating the alert by placing a call or sending a text message.
31. The method according to claim 18 , further comprising performing the calibration procedure by:
receiving a calibration speech signal representing other speech uttered by the subject,
receiving an airflow-rate signal representing measured flow rates of air exhaled by the subject while uttering the other speech, and
using the calibration speech signal and the airflow-rate signal, computing the function.
32. The method according to claim 31 , further comprising measuring the airflow-rate signal using a pneumotach.
33. The method according to claim 32 , further comprising recording the calibration speech signal using a sound-capturing unit disposed inside of the pneumotach.
34. A computer software product comprising a tangible non-transitory computer-readable medium in which program instructions are stored, which instructions, when read by one or more processors, cause the processors to cooperatively carry out a process that includes:
receiving, via an audio-receiving device, a speech signal that represents speech uttered by a subject,
dividing at least part of the speech signal into multiple frames representing respective portions of the speech,
applying a function, which was computed during a calibration procedure, to one or more features of each of the frames so as to compute respective estimated volumetric flow rates of air exhaled by the subject while uttering the portions of the speech,
computing, based on the estimated volumetric flow rates, an estimated total volume of air exhaled by the subject while uttering the portions of the speech, and
in response to the estimated total volume of air, communicating an alert from a server to the audio-receiving device.