Phonopneumograph system
AssigneeKarmel Medical Acoustic Technologies Ltd.
InventorNoam Gavriely
Life of this patent
Estimating the term…
Abstract
A phonopneumograph system for analyzing breath sounds includes a plurality of breath related sensors placed around the respiratory system of a patient for measuring breath related activity and a breath analyzer. The breath analyzer matches the breath sound data produced by the breath related sensors to a plurality of breath sound templates each of which parametrize one type of breath sound and determines the presence of regular and/or adventitious breath sounds only when the breath sound data matches, within predetermined goodness of fit criteria, one or more of the breath sound templates.
Claims90
A phonopneumograph system for analyzing breath sounds produced by a respiratory system within a patient's body, the system comprising:
- at least one breath sounds sensor adapted to receive breath sounds conducted through a patient's body and produce at least one breath sounds signal responsive to thc breath sounds;
- and
- a breath sounds analyzer which receives said breath sounds signal and matches said signal or data derived from said signal to at least one breath sounds template, at least one of said templates parametrizing regular chest wall breath sounds,
- wherein the template for regular chest wall breath sounds is a frequency domain curve of a low pass filter, wherein the parameters are an amplitude, cutoff frequency and slope of the low pass filter.
A phonopneumograph system for analyzing breath sounds produced by a respiratory system within a patient's body, the system comprising:
- at least one breath sounds sensor adapted to receive breath sounds conducted through a patient's body and produce at least one breath sounds signal responsive to the breath sounds;
- and
- a breath sounds analyzer which receives said breath sounds signal and matches said signal or data derived from said signal to at least one breath sounds template, at least one of said templates parametrizing regular tracheal breath sounds,
- wherein the template for regular tracheal breath sound is a frequency domain curve of an ensemble of second order systems wherein the parameters are a set of amplitude coefficients, a set of damping coefficients and a set of resonance frequencies.
A phonopneumograph system for analyzing breath sounds produced by a respiratory system within a patient's body, the system comprising:
- at least one breath sounds sensor adapted to receive breath sounds conducted through a patient's body and produce at least one breath sounds signal responsive to the breath sounds;
- and
- a breath sounds analyzer which receives said breath sounds signal and matches said signal or data derived from said signal to at least one breath sounds template, at least one of said templates parametrizing a wheeze;
- wherein the template for a wheeze is a narrow frequency domain peak whose width at half height spreads less than 32 Hz to either side of a center frequency of said narrow peak and has less than three harmonics, said narrow peak occurring within at least three time segments spanning between 150 and 2500 ms wherein said frequency of said narrow peak varies less than a predetermined amount within said three segments.
A phonopneumograph system for analyzing breath sounds produced by a respiratory system within a patient's body, the system comprising:
- at least one breath sounds sensor adapted to receive breath sounds conducted through a patient's body and produce at least one breath sounds signal responsive to the breath sounds;
- and
- a breath sounds analyzer which receives said breath sounds signals and matches said signals or data derived from said signals to at least one breath sounds template, at least one of said templates parametrizing a rhonchus;
- wherein the template for a rhonchus is a repetitive sound in said breath sound data having a series of generally evenly spaced time and frequency domain peaks wherein the sound has a significant lack of time and frequency correlation with ambient noise.
A phonopneumograph system for analyzing breath sounds produced by a respiratory system within a patient's body, the system comprising:
- at least one breath sounds sensor adapted to receive breath sounds conducted through a patient's body and produce at least one breath sounds signal responsive to the breath sounds;
- and
- a breath sounds analyzer which receives said breath sounds signal and matches said signal or data derived from said signal to at least one breath sounds template, at least one of said templates parametrizing a snore;
- wherein the template for a snore is a repetitive sound, that is significantly correlated with ambient noise, and has generally evenly spaced time domain sound structures and evenly spaced frequency domain peaks, wherein an average time spacing between sound structures is generally equal to one divided by an average of the spacing between frequency peaks.
A phonopneumograph system for analyzing breath sounds produced by a respiratory system within a patient's body, the system comprising:
- at least one breath sounds sensor adapted to receive breath sounds conducted through a patient's body and produce at least one breath sounds signal responsive to the breath sounds;
- and
- a breath sounds analyzer which receives said breath sounds signal and matches said signal or data derived from said signal to at least one breath sounds template, at least one of said templates parametrizing a cough;
- wherein the template for a cough is a sound occurring during expiration that has a significant correlation with ambient noise and which endures 0.2-3 seconds and has a double hump time domain envelope and a relatively flat frequency spectrum.
A phonopneumograph system for analyzing breath sounds produced by a respiratory system within a patient's body, the system comprising:
- at least one breath sounds sensor adapted to receive breath sounds conducted through a patient's body and produce at least one breath sounds signal responsive to the breath sounds;
- and
- a breath sounds analyzer which receives said breath sounds signals and matches said signals or data derived from said signals to at least one breath sounds template, at least one of said templates parametrizing a crackle;
- wherein the template for a crackle is a curve, whose onset point begins as an abrupt change in said breath sound data, which generally matches:
- y=A*B(t)*C(t)
- where y is a breath sound signal beginning at said onset point, t begins at said onset point, A is an amplitude parameter, B(t) is an envelope function and C(t) is an oscillatory function.
A phonopneumograph system for analyzing breath sounds produced by a respiratory system within a patient's body, the system comprising:
- at least one breath sounds sensor adapted to receive breath sounds conducted through a patient's body and produce at least one breath sounds signal responsive to the breath sounds;
- and
- a ambient noise microphone, adapted to produce an ambient noise signal responsive to ambient noise;
- an ambient noise level detector that quantifies the ambient noise signal as representing loud ambient noise or a low level of ambient noise;
- a loud noise analyzer that analyzes the breath sound signal to determine whether at least one of snores and coughs are present when said ambient noise is quantified as loud ambient noise;
- and
- a breath analyzer that analyzes the breath sound signal to determine breathing activity and to detect adventitious breath sounds in said breath sound signals when said ambient noise is quantified as low level noise.
A breath sounds for monitoring breath sounds produced by a respiratory system within a patient's body, the system comprising:
- at least one breath sounds sensor adapted to receive breath sounds conducted through a patient's body and produce at least one breath sounds signal responsive to the breath sounds;
- and
- a breath sounds analyzer which receives said at least one breath sounds signal and matches the breath sound signal or data produced from the breath sound signal to a plurality of breath sound templates each of which parametrizes one type of breath sound and the presence of regular or adventitious breath sounds only when said breath sound data or signal matches, within predetermined goodness of fit criteria, one or more of said breath sound templates;
- and
- a log unit which logs template parameters of at least said detected adventitious sounds.
A breath sounds meter for measuring cyclic breath sounds produced by a respiratory system within a patient's body, the system comprising;
- at least one breath sounds sensor adapted to receive breath sounds conducted through a patient's body and produce at least one breath sounds signal responsive to the breath sounds;
- and
- a breath sounds analyzer that receives said breath sounds signal and determines, responsive to the signal or data derived from the signal, regular breath sounds or adventitious breath sounds;
- and
- a display that indicates at least regular breathing and when in the respiratory cycle each of said adventitious sounds occurs and which displays at least a parametric description of said breath sounds.
A phonopneumographic system comprising:
- a plurality of piezoelectric sensors adapted to be placed around a patient's respiratory system for measuring breath related activity;
- and
- a verification unit connected to said sensors for individually activating each sensor to produce a sound and for each sound produced, measuring said sound with the non-activated ones of said sensors and, in a verification mode, comparing the measured sounds with those received during a training mode.
A hardware verification method for a system having a plurality of piezoelectric transducers placed around an object, the method comprising:
- in a training mode:
- individually activating at least some of said transducers to produce at least one sound;
- and
- for each sound produced, measuring said sound with non-activated ones of said transducers;
- in a verification mode:
- individually activating at least some of said transducers to produce a at least one sound;
- for each sound produced, measuring said sound with non-activated ones of said transducers;
- and
- identifying malfunctioning sensors by comparing the measured sounds with those received during said training mode.
A method for detecting a wheeze in breath sound data, the method comprising:
- segmenting said breath sound data into segments;
- per segment:
- generating the spectrum of said breath sound segment;
- removing noise and regular breath sound spectrum patterns from said breath sound spectrum, thereby to produce a non-regular breath sound spectrum;
- detecting narrow peaks within said non-regular breath sound spectrum;
- per a small group of segments:
- defining a potential wheeze if narrow peaks exist in consecutive segments and if said narrow peaks are located within a predetermined small frequency range across said consecutive segments;
- defining a wheeze if the narrow peaks of said potential wheeze have less than three harmonics each.
A method of detecting a rhonchus in breath sound data, the method comprising:
- segmenting said breath sound data into segments;
- per segment, detecting narrow peaks within a spectrum of said segment;
- per a small group of segments:
- defining a potential rhonchus if narrow peaks exist in consecutive segments and if said narrow peaks are located within a predetermined small frequency range across said consecutive segments;
- if there are more than two harmonics in each segment,
- generating a frequency transfer function between said breath sound data and measured ambient noise of a space where said breath sound data was gathered;
- determining a coherence graph of the transfer function;
- and
- defining each narrow peak as a ronchus if the frequency range of high coherence of said coherence graph does not correspond to the frequency range of said narrow peaks.
A method of detecting a cough in breath sound data, the method comprising:
- generating an amplitude spectrum of breath sound data and of ambient noise data;
- generating a transfer function between said amplitude spectra of said breath sound and ambient noise data;
- determining a coherence graph of said transfer function;
- finding peaks in said breath sound and ambient noise data;
- generating an envelope of said breath sound data, having a duration and determining the duration of said envelope;
- reviewing said breath sound and ambient noise data and identifying a cough if said breath sound and ambient noise fulfill all of the following conditions:
- a) the breath sound is an expiration sound;
- b) said peaks of said breath sound data generally coincide with said peaks of said ambient noise data;
- c) said coherence graph has a significant portion which has a high coherence level;
- d) said envelope of said breath sound data has a double hump shape;
- and
- e) the duration of said envelope is a predetermined length of time.
A method of detecting crackles in breath sound data, the method comprising the steps of:
- finding locations of abrupt changes in said breath sound data;
- for each abrupt change,
- finding a beginning point of said abrupt change;
- attempting to match said breath sound data following said beginning point to:
- y=A*B(t)*C(t) (17)
- where y is said breath sound data from said beginning point, t begins at said beginning point, A is an amplitude parameter, B(t) is an envelope function and C(t) is an oscillatory function;
- and
- identifying a crackle if said breath sound data matches said curve.
A method of determining a breathing state of a patient, the method comprising:
- determining an inspiration/expiration phase of a breath from chest movement data and defining a breath variable therefrom;
- acquiring tracheal breath sound data from a subject;
- if the tracheal breath sound data are significant and if external noise is low:
- determining if the tracheal breath sound data has a generally normal shape;
- and
- if so, generating breath flow data from said tracheal breath sound data and from said breath phase variable;
- otherwise:
- determining if the lack of flow indicates apnea and, if so, setting an apnea alarm.
A method for analyzing breath data, the method comprising:
- generating a spectrum of a current segment;
- determining if the current segment of data is a background segment, representing background noise, or a breath segment representing breath sounds;
- averaging spectra of a plurality of segments of each type to produce an average background spectrum and an average breath sound spectrum;
- for said average breath sound spectrum, subtracting said average background spectrum therefrom to produce a relatively noiseless breath spectrum;
- fitting said relatively noiseless breath spectrum to a predetermined normal curve and determining a quality of fit;
- activating an irregular breath sounds detector if said quality of fit is poor.
A method of detecting a snore in segments of breath sound data, said segments having a beginning and an end, the method comprising:
- determining that a current segment is an inspiratory segment;
- identifying amplitude peaks in said inspiratory segment and determining an average peak-to-peak time delta_t;
- identifying at least three frequency peaks within a range of frequencies in a power spectrum of said breath sound data which are significantly large and determining an average peak-to-peak frequency delta_f;
- generating a frequency transfer function between said breath sound data and measured ambient noise where said breath sound data was gathered;
- determining a coherence graph of said transfer function;
- and
- identifying a snore if delta_t multiplied by delta_f is close to 1 and if a frequency range of high coherence of said coherence graph includes said frequency peaks.
A phonopneumograph system for analyzing breath sounds produced by a respiratory system within a patient's body, the system comprising:
- at least one breath sounds sensor adapted to receive breath sounds conducted through a patient's body and produce at least one breath sounds signal responsive to the breath sounds;
- and
- a ambient noise microphone, adapted to produce an ambient noise signal responsive to ambient noise when placed near said patient;
- an ambient noise level detector that quantifies the ambient noise signal as representing loud ambient noise or low level ambient noise;
- and
- a breath analyzer that analyzes the breath sound signal to determine the presence of regular breathing activity or adventitious breath sounds including one or more of a wheeze, a rhonchus or a cackle in said breath sound signals only when said ambient noise is quantified as low level noise.
A phonopneumograph for analyzing breath sounds, produced by a respiratory system the phonopneumograph comprising:
- at least one breath related sensor adapted to be placed adjacent the respiratory system of a patient, the sensor measuring breath related activity, said at least one sensor producing breath sound data;
- a breath analyzer which matches said breath sound data to a plurality of breath sound templates each of which is characteristic of and parametrizes one type of breath sound and which produces a signal indicative of the presence of a particular regular or adventitious breath sounds only when said breath sound data matches, within predetermined goodness of fit criteria, one or more of said breath sound templates;
- wherein the sound templates comprise particular standard time or frequency domain patterns characteristic of a particular breath sound and wherein the parameters define particular variations in said pattern.
Dependencies are inferred from the claim text.