JP2005066044A - Respiratory sound data processor and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a respiratory sound data processor which can acquire respiratory sound data wherein a flow rate and/or a flow velocity (or a respiratory time and/or a respiratory sound strength) are normalized without using a flowmeter, can support an effective diagnosis by analyzing the respiratory sound data, and is compact and has a simple constitution, and a program for analyzing the respiratory sound data by controlling the respiratory sound data processor. <P>SOLUTION: The respiratory sound data for one respiration is acquired by a one-respiration sound extracting means 131, and the respiratory time and the respiratory sound strength are calculated from the acquired respiratory sound data for one respiration by a one-respiration sound normalizing means 132. The flow rate and/or the flow velocity for one respiration are calculated based on the respiration time and the respiratory sound strength, and whether the flow rate and/or the flow velocity are normalized or not is judged. By a subsidiary sound detecting means 133, the presence/absence or the like of subsidiary sounds contained in the respiration sound is detected based on the respiratory sound data of the normalized flow rate and/or the flow velocity. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a respiratory sound data processing apparatus and program for analyzing and processing respiratory sound data representing respiratory sounds.

  Conventionally, clinicians have made various diagnoses based on breath sounds directly heard through a stethoscope. Here, “respiratory sounds” is a general term for sounds that occur with respiratory movement and can be auscultated in the chest. And “adventitious sounds” which are abnormal sounds. Hereinafter, “breathing sound” is used for the above “respiratory sounds”.

  Recently, various techniques for storing respiratory sounds as digital data (hereinafter referred to as respiratory sound data) and effectively utilizing them for later diagnosis have been devised (see, for example, Patent Documents 1 and 2). In order to analyze such breathing sound data, the flow rate and / or flow rate of one breath, or the breathing time and / or breathing sound intensity of one breath are kept constant (hereinafter referred to as normalization). Is preferred. Conventionally, a flow meter has been generally used as a device for standardizing the flow rate and / or flow velocity.

In this way, accurate analysis can be realized by using the respiratory sound data in which the flow rate and / or flow rate of one breath (or the breathing time and / or the intensity of the breathing sound) is standardized.
Japanese Patent Laid-Open No. 5-73093 Japanese translation of PCT publication No. 2002-511295

  However, the technique for standardizing the flow rate and / or flow rate of one breath using the flow meter as described above has the following problems. That is, a respiratory sound data processing apparatus for analyzing respiratory sound data using a flow meter is large and complicated by including the flow meter. In addition, a respiratory sound data processing apparatus that can measure the flow rate and / or flow velocity (or respiratory time and / or respiratory sound intensity) of a breath without using a flow meter has not been devised.

  An object of the present invention is to obtain respiratory sound data in which the flow rate and / or flow velocity (or breathing time and / or respiratory sound intensity) of one breath is standardized without using a flow meter. A respiratory sound data processing device having a small and simple configuration that enables effective diagnosis support by analyzing the above, and a program for controlling the respiratory sound data processing device to analyze respiratory sound data That is.

In order to solve the above-mentioned problem, the invention described in claim 1
A respiratory sound data processing apparatus for analyzing and processing respiratory sound data representing respiratory sounds,
One breath sound extraction means for extracting data for one breath from the breath sound data;
A breathing sound normalization means for calculating a breathing time and breathing sound intensity from the breathing sound data for one breath and normalizing the breathing sound data for the breathing sound based on the calculated breathing time and breathing sound intensity When,
It is provided with.

Further, as in the invention according to claim 2, in the invention according to claim 1,
Breathing sound signal acquisition means for acquiring a breathing sound signal representing the breathing sound;
A / D conversion means for digitizing the respiratory sound signal acquired by the respiratory sound signal acquisition means into respiratory sound data;
Further comprising
The breathing sound extraction means includes
It is preferable to extract data for one breath from breathing sound data digitized by the A / D conversion means.

Further, in the invention according to claim 1 or 2, as in the invention according to claim 3,
The breathing sound extraction means includes
When two consecutive timings when the magnitude of the respiratory sound related to the respiratory sound data acquired via the respiratory sound signal acquisition unit exceeds a preset threshold are detected, the two detected timings are used. It is preferable to identify one breath and extract data for one breath from the breathing sound data.

Further, in the invention according to claim 1 or 2, as in the invention according to claim 4,
The breathing sound extraction means includes
It is preferable to calculate a correlation value between the respiratory sound data and preset sample respiratory sound data for one breath, and extract data for one breath from the respiratory sound data based on the calculated correlation value. .

Further, as in the invention according to claim 5, in the invention according to claim 4,
The breathing sound extraction means includes
The correlation value is calculated using a plurality of preset sample respiratory sound data having different breathing time and / or respiratory sound intensity for one breath, and data for one breath is calculated based on the calculated plurality of correlation values. Is preferably extracted.

Furthermore, as in the invention according to claim 6, in the invention according to claim 1,
Comprising first to nth (n is a natural number of 2 or more) respiratory sound signal acquisition means for acquiring a respiratory sound signal representing a respiratory sound;
The A / D conversion means includes
Preferably, each of the plurality of respiratory sound signals acquired by the first to nth respiratory sound signal acquisition means is digitized into respiratory sound data.

Furthermore, as in the invention according to claim 7, in the invention according to claim 6,
The breathing sound extraction means includes
The timing when the magnitude of the breathing sound related to the breathing sound data acquired via the first breathing sound signal acquisition unit exceeds a predetermined threshold set in advance according to the first breathing sound signal acquisition unit. After the detection, the second to nth respiratory sound signal acquisition means has a magnitude of the respiratory sound related to the respiratory sound data acquired through at least one of the second to nth respiratory sound signal acquisition means. Each of which exceeds a predetermined threshold value set in advance according to each of the first and second respiratory sounds, and the magnitude of the respiratory sound related to the respiratory sound data acquired through the first respiratory sound signal acquisition means is the first respiratory sound. When the timing when the signal acquisition means exceeds the threshold again is detected, it is preferable to extract data for one breath from the breathing sound data using the two timings.

Further, as in the invention according to claim 8, in the invention according to claim 1,
The breathing sound extraction means includes
Each start timing of exhalation and inspiration is alternately generated and notified every predetermined time interval, and breathing sound data acquired according to the start timing of expiration is used as breathing sound data corresponding to expiration, and the start timing of inspiration It is preferable to acquire the respiratory sound data for one breath as the respiratory sound data corresponding to the inspiration.

Further, as in the invention according to claim 9, in the invention according to any one of claims 1 to 8,
The breathing sound normalization means includes:
Flow rate / flow velocity calculation means for calculating a flow rate and / or flow velocity for one breath based on the breathing time and breathing sound intensity for one breath is further provided, and one breathing sound is calculated based on the calculated flow rate and / or flow velocity. It is preferable to normalize the minute breathing sound data.

Furthermore, as in the invention according to claim 10, in the invention according to claim 9,
The breathing sound normalization means includes:
The flow rate for one breath is preferably calculated by multiplying the breathing time for one breath by the respiratory sound intensity.

Further, in the invention according to claim 9 or 10, as in the invention according to claim 11,
The breathing sound normalization means includes:
It is preferable to calculate the flow rate for one breath by dividing the respiratory sound intensity for one breath by the breathing time.

Furthermore, as in the invention according to claim 12, in the invention according to any one of claims 9 to 11,
The breathing sound normalization means includes:
The flow rate and / or flow rate is preferably corrected using personal information for each subject.

Furthermore, in the invention according to claim 12, like the invention according to claim 13,
The breathing sound normalization means includes:
The flow rate and / or flow rate is preferably corrected using personal information including at least one of the height, weight, sex, obesity level, medical history, and vital capacity of the subject.

Further, as in the invention according to claim 14, in the invention according to any one of claims 9 to 13,
The breathing sound normalization means includes:
It is determined whether or not the flow rate and / or flow velocity calculated by the flow rate / velocity calculation means matches a preset standard and flow rate and / or flow velocity. Is preferably notified.

Further, as in the invention according to claim 15, in the invention according to any one of claims 1 to 8,
The breathing sound normalization means includes:
Breathing sound intensity normalization means for normalizing the breathing sound intensity related to breathing sound data for one breath;
Breathing time normalization means for normalizing the breathing time related to breathing sound data for one breath;
With
It is preferable that the breathing sound data for one breath is standardized using the breathing sound intensity normalization means and the breathing time normalization means.

Further, as in the invention according to claim 16, in the invention according to claim 15,
The breathing sound intensity normalization means is
The respiratory sound data for one breath is separated into low-frequency respiratory sound data whose respiratory sound frequency is equal to or lower than a predetermined reference frequency and high-frequency respiratory sound data exceeding the reference frequency, and the respiratory sound intensity of the low-frequency respiratory sound data is determined. It is preferable to normalize and standardize the respiratory sound intensity related to the respiratory sound data for one breath by synthesizing the normalized low-frequency respiratory sound data and the high-frequency respiratory sound data.

Further, as in the invention according to claim 17, in the invention according to claim 15,
The breathing sound intensity normalization means is
Among the low-frequency respiratory sound data whose respiratory sound frequency is preliminarily separated from the respiratory sound data for one breath and the high-frequency respiratory sound data whose frequency exceeds the reference frequency, the respiratory sound of the low-frequency respiratory sound data Preferably, the intensity is normalized, and the normalized low-frequency respiratory sound data is combined with the high-frequency respiratory sound data to normalize the respiratory sound intensity related to the respiratory sound data for one breath.

Further, as in the invention according to claim 18, in the invention according to any one of claims 15 to 17,
The breathing time normalization means includes
It is preferable to normalize the breathing time using the dynamic time warping method for the breathing sound data for one breath.

Further, as in the invention described in claim 19, in the invention described in any one of claims 1 to 18,
It is preferable to further include sub noise detection means for detecting sub noise included in the breathing sound from the normalized breathing sound data for one breath.

Furthermore, as in the invention according to claim 20, in the invention according to claim 19,
The sub-noise detection means includes
The spectrum information related to the breathing sound data is obtained by performing FFT processing on the normalized breathing sound data for one breath, and sub-noise is detected by applying a hidden Markov process to the spectrum information. Is preferred.

Furthermore, as in the invention according to claim 21, in the invention according to claim 19,
The sub-noise detection means includes
It is preferable to calculate a correlation value between two normalized breathing sound data for one breath acquired from the same subject at two different times, and detect sub-noise based on the calculated correlation value.

Furthermore, as in the invention according to claim 22, in the invention according to claim 19,
The sub-noise detection means includes
A correlation value between the standardized breathing sound data for one breath and the sampled breathing sound data not including the secondary noise is calculated, and the secondary noise is detected based on the calculated correlation value. Is preferred.

Furthermore, as in the invention described in claim 23, in the invention described in any one of claims 4, 5, 9-22,
The correlation value is preferably a cross-correlation value or a residual sum of squares.

In order to solve the above problem, the invention according to claim 24 provides
A computer for analyzing and processing respiratory sound data representing respiratory sounds,
A function of extracting data for one breath from the breathing sound data;
A function for calculating a breathing time and a breathing sound intensity from the breathing sound data for the one breath, and normalizing the breathing sound data for the one breathing sound based on the calculated breathing time and the breathing sound intensity;
Is realized.

Furthermore, in the invention according to claim 24, as in the invention according to claim 25,
In the computer,
When two consecutive timings when the magnitude of the respiratory sound related to the respiratory sound data exceeds a preset threshold are detected, one respiratory is identified using the two detected timings, and the respiratory sound data It is preferable to further realize a function of extracting data for one breath from the data.

Furthermore, in the invention according to claim 24, as in the invention according to claim 26,
In the computer,
A function of calculating a correlation value between the respiratory sound data and preset sample respiratory sound data for one breath, and further extracting data for one breath from the respiratory sound data based on the calculated correlation value Preferably it is realized.

Furthermore, in the invention according to claim 24, as in the invention according to claim 27,
In the computer,
The correlation value is calculated using a plurality of preset sample respiratory sound data having different breathing time and / or respiratory sound intensity for one breath, and data for one breath is calculated based on the calculated plurality of correlation values. It is preferable to further realize the function of extracting.

Furthermore, in the invention according to claim 24, as in the invention according to claim 28,
In the computer,
Among the respiratory sound data acquired via the first to nth respiratory sound signal acquisition means, the magnitude of the respiratory sound related to the respiratory sound data acquired via the first respiratory sound signal acquisition means Obtained via at least one of the second to nth respiratory sound signal acquisition means after detecting a timing when a predetermined threshold value preset in accordance with the first respiratory sound signal acquisition means is exceeded. The magnitude of the breathing sound related to the breathing sound data exceeds each of the predetermined threshold values set in advance according to each of the second to nth breathing sound signal acquisition means, and then the first breathing sound signal acquisition means When the timing at which the magnitude of the breathing sound related to the breathing sound data acquired via the above again exceeds the threshold value of the first breathing sound signal acquisition means is detected, the breathing is performed using the two detected timings. One breath data from sound data Preferably further to realize the function of extracting.

Further, as in the invention described in claim 29, in the invention described in claim 24,
In the computer,
Each start timing of exhalation and inspiration is alternately generated and notified every predetermined time interval, and breathing sound data acquired according to the start timing of expiration is used as breathing sound data corresponding to expiration, and the start timing of inspiration It is preferable to further realize a function of acquiring the respiratory sound data for one breath as the respiratory sound data corresponding to the inspiration.

Further, as in the invention described in claim 30, in the invention described in any one of claims 24-29,
In the computer,
Flow rate / flow velocity calculation means for calculating a flow rate and / or flow velocity for one breath based on the breathing time and breathing sound intensity for one breath is further provided, and one breathing sound is calculated based on the calculated flow rate and / or flow velocity. It is preferable to further realize a function of normalizing minute respiratory sound data.

Further, as in the invention of claim 31, in the invention of claim 30,
In the computer,
It is preferable to further realize a function of calculating the flow rate for one breath by multiplying the breathing sound intensity by the breathing time for one breath.

Furthermore, in the invention according to claim 30 or 31, as in the invention according to claim 32,
In the computer,
It is preferable to further realize a function of calculating the flow rate for one breath by dividing the respiratory sound intensity for one breath by the breathing time.

Further, as in the invention described in claim 33, in the invention described in any one of claims 30 to 32,
In the computer,
It is preferable to further realize a function of correcting the flow rate and / or flow rate using personal information for each subject.

Further, in the invention according to claim 33, as in the invention according to claim 34,
In the computer,
It is preferable to further realize the function of correcting the flow rate and / or flow rate using personal information including at least one of the height, weight, sex, obesity level, medical history, and vital capacity of the subject.

Further, as in the invention described in claim 35, in the invention described in any one of claims 30 to 34,
In the computer,
It is determined whether or not the flow rate and / or flow velocity calculated by the flow rate / velocity calculation means matches a preset standard and flow rate and / or flow velocity. It is preferable to further realize the function of informing the user.

Further, as in the invention described in claim 36, in the invention described in any one of claims 24 to 29,
In the computer,
A function to normalize the breathing sound intensity associated with breathing sound data for one breath;
A function for normalizing the breathing time associated with breathing sound data for one breath;
Is preferably realized.

Furthermore, in the invention of claim 36, as in the invention of claim 37,
In the computer,
Respiratory sound intensity of the low-frequency respiratory sound data is standardized by classifying the respiratory sound data for one breath into low-frequency respiratory sound data whose respiratory sound frequency is equal to or lower than a predetermined reference frequency and high-frequency respiratory sound data exceeding the reference frequency. It is preferable to further realize a function of normalizing the respiratory sound intensity related to the respiratory sound data for one breath by synthesizing the normalized low frequency respiratory sound data and the high frequency respiratory sound data.

Furthermore, in the invention of claim 36, as in the invention of claim 38,
In the computer,
The respiratory sound intensity of the low-frequency respiratory sound data among the low-frequency respiratory sound data whose respiratory sound frequency is preliminarily separated from the respiratory sound data for one breath and the high-frequency respiratory sound data that exceeds the reference frequency It is preferable to further realize a function of normalizing the respiratory sound intensity related to the respiratory sound data for one breath by combining the normalized low frequency respiratory sound data with the high frequency respiratory sound data.

Furthermore, in the invention according to any one of claims 36 to 38, as in the invention according to claim 39,
In the computer,
It is preferable to further realize the function of normalizing the breathing time using the dynamic time warping method for the breathing sound data for one breath.

Furthermore, as in the invention described in claim 40, in the invention described in any one of claims 24-39,
In the computer,
It is preferable to further realize a function of detecting secondary noise included in the breathing sound from the normalized breathing sound data for one breath.

Further, as in the invention of claim 41, in the invention of claim 40,
In the computer,
A function of performing FFT processing on the normalized breathing sound data for one breath to obtain spectrum information related to the breathing sound data, and detecting sub-noise by applying a hidden Markov process to the spectrum information Is preferably realized.

Furthermore, as in the invention of claim 42, in the invention of claim 40,
In the computer,
A function to calculate the correlation value between two standardized breathing sound data for one breath acquired from the same subject at two different times and to detect sub-noise based on the calculated correlation value is further realized. It is preferable to do so.

Furthermore, as in the invention of claim 43, in the invention of claim 40,
In the computer,
A correlation value between the standardized breathing sound data for one breath and the sampled breathing sound data not including the secondary noise is calculated, and the secondary noise is detected based on the calculated correlation value. It is preferable to further realize the function.

Further, as in the invention according to claim 44, in the invention according to any one of claims 26, 27, and 30 to 43,
The correlation value is preferably a cross-correlation value or a residual sum of squares.

Furthermore, in the invention according to claim 21 or 22, as in the invention according to claim 45,
It is preferable to further include output means for displaying each of the respiratory sound waveforms represented by the two respiratory sound data and / or reproducing the two respiratory sounds.

Furthermore, in the invention of claim 45, like the invention of claim 46,
The output means includes
It is preferable that the respiratory sound wave forms represented by the two respiratory sound data are displayed side by side.

Further, in the invention according to claim 42 or 43, as in the invention according to claim 47,
In the computer,
It is preferable to further realize a function of displaying each of the respiratory sound waveforms represented by the two respiratory sound data and / or reproducing the two respiratory sounds.

Furthermore, as in the invention according to claim 48, in the invention according to claim 47,
In the computer,
It is preferable that the respiratory sound wave forms represented by the two respiratory sound data are displayed side by side.

  Therefore, the flow rate and / or flow rate for one breath, or the breathing time and breathing sound intensity for one breath can be determined based on only the breathing sound data acquired through the breathing sound signal acquisition means without using a flow meter. Respiratory sound data with standardized and standardized flow rate and / or flow velocity (or breathing time and respiratory sound intensity for one breath) is analyzed (detection of secondary noise, etc.). It is possible to realize a respiratory sound data processing apparatus having a small and simple configuration capable of providing simple diagnosis support.

A respiratory sound data processing apparatus 100 to which the present invention is applied will be described in detail with reference to the drawings.
As shown in FIG. 1, the respiratory sound data processing apparatus 100 includes a respiratory sound signal acquisition unit 11, an A / D conversion unit 12, a data processing unit 13, and an output unit 14.

  The respiratory sound signal acquisition unit 11 includes one auscultation microphone (not shown), and the respiratory sound collected by the microphone is converted into an analog respiratory sound signal (that is, a respiratory sound waveform signal). The data is converted and output to the A / D conversion means 12. The A / D conversion means 12 digitizes the respiratory sound signal input from the respiratory sound signal acquisition means 11 based on a predetermined sampling frequency and converts it into respiratory sound data. Output to the breathing sound extraction means 131.

  The data processing means 13 includes a breathing sound extraction means 131, a breathing sound normalization means 132, and a sub-noise detection means 133, and the breathing sound normalization means 132 includes a flow rate / flow velocity calculation means 1321.

  The data processing means 13 executes the breathing sound data processing shown in the flowcharts of FIGS. 2 and 3 using the one breathing sound extraction means 131 to the auxiliary noise detection means 133. Further, the data processing means 13 outputs image information and sound information necessary for this breathing sound data processing, or image information and sound information representing the processing result to the output means 14.

  Note that the breathing sound extraction unit 131 to the sub-noise detection unit 133 may be hardware, or may represent functions included in a program executed by the data processing unit 13. Here, the program is a program for performing the respiratory sound data processing shown in the flowcharts of FIGS. 2 and 3 and stored in advance in a built-in memory (not shown) of the data processing means 13. To do.

  Details of the operations of the one-breathing sound extraction means 131 to the auxiliary noise detection means 133 will be described later based on the flowcharts of FIGS.

  The output unit 14 includes a display device such as an LCD (Liquid Crystal Display), a speaker (both not shown), and the like, and displays various image information and sound information input from the data processing unit 13 on the display device. Or emit sound from the speaker.

  Next, the operation of the respiratory sound data processing apparatus 100 will be described with reference to FIGS.

First, the respiratory sound data processing performed by the data processing means 13 will be described in detail with reference to FIG.
First, the breathing sound extraction unit 131 executes the breathing sound data processing shown in the flowchart of FIG. 3 to be described later, and extracts the breathing sound data for one breath from the breathing sound data input from the A / D conversion unit 12. (Step S10). The breathing sound extraction unit 131 further outputs the extracted breathing sound data for the breathing to the flow rate / flow velocity calculation unit 1321 of the breathing sound normalization unit 132.

  After step S10, the flow rate / flow velocity calculation means 1321 calculates the breathing time and breathing sound intensity for one breath based on the breathing sound data for one breath input from the one breathing sound extraction means 131, and is shown below. Based on Formula 1, the flow rate and / or flow velocity relating to the one breath is calculated (step S11). Hereinafter, the respiratory sound intensity is, for example, the average intensity or the maximum intensity of the respiratory sound in one breath.

  After step S <b> 11, the breathing sound normalization unit 132 uses the flow rate and / or flow rate for which the flow rate and / or flow rate for one breath calculated by the flow rate / flow rate calculation unit 1321 is a preset standard for normalization. (I.e., whether the flow rate and / or flow velocity are normalized) (step S12).

  In step S12, the one-breathing sound normalization unit 132 determines that the calculated flow rate and / or flow velocity does not match the flow rate and / or flow velocity that is the standard for the normalization (when it is not standardized). (Step S12; No), the image information or the sound information for notifying the subject to that effect and instructing the patient to breathe at a flow rate and / or a flow rate that meets the above-mentioned criteria is output to the output means 14 (Step S13). ) When the calculated flow rate and / or flow velocity matches the standardized flow rate and / or flow velocity (when standardized) (step S12; Yes), The respiratory sound data is output to the secondary noise detection means 133.

  After the branch of step S12; Yes, the sub-noise detecting means 133, based on the respiratory sound data for one breath with the flow rate and / or flow velocity normalized, input from the one breath sound normalizing means 132, The secondary noise included in the breathing sound is detected (step S14).

  That is, the sub-noise detection means 133 separates the breathing sound data into two breathing sound data corresponding to expiration and inspiration, and separately for the breathing sound data corresponding to expiration and the breathing sound data corresponding to inspiration. The sub-noise is detected. Here, a method for separating the breathing sound data for one breath into breathing sound data corresponding to exhalation and breathing sound data corresponding to inspiration is, for example, approximately dividing the breathing sound data for one breath in the time direction. There is a method in which the breathing sound data includes data indicating the timing at which the breathing sound extraction means 131 changes from expiration to inspiration (or inspiration to expiration) in advance. For example, the data indicating the timing is manually input by an operator such as a clinician or automatically generated based on a difference in waveforms of exhalation and inspiration in the respiratory sound data.

  Next, with reference to FIG. 3, the breathing sound data processing performed by the breathing sound extraction means 131 will be described in detail.

  It is assumed that the first respiratory sound signal acquisition unit 111 is disposed in a predetermined part (for example, near the bronchus) of the subject's body.

  The breathing sound extraction unit 131 acquires the breathing sound data via the breathing sound signal acquisition unit 11 and the A / D conversion unit 12 (step S20), and the magnitude of the breathing sound represented by the breathing sound data is set in advance. It is monitored whether or not the threshold value Th1 is exceeded (that is, whether or not one breath is started) (step S21).

  When the magnitude of the breathing sound exceeds the threshold Th1 (step S21; Yes), the breathing sound extraction means 131 detects the timing T1 at this time (that is, at the start of one breathing) (step S22). Data indicating the detected timing T1 is temporarily stored in the internal memory.

  After step S22, the breathing sound extraction means 131 monitors whether or not the magnitude of the breathing sound has become smaller than a preset threshold value Th2 (that is, whether or not a breathing is in progress). (Step S23) When the magnitude of the breathing sound becomes smaller than the threshold Th2 (Step S23; Yes), the timing T2 at this time is detected (Step S24). Data indicating the detected timing T2 is temporarily stored in the built-in memory.

  After step S24, the breathing sound extraction means 131 monitors whether or not the magnitude of the breathing sound has again exceeded the threshold value Th1 (that is, whether or not the breathing has ended) (step S25). When the volume of the breathing sound exceeds the threshold Th1 again (step S25; Yes), the timing T1 ′ at this time (that is, at the end of one breath) is detected (step S26).

After step S26, the breathing sound extraction unit 131 identifies one breath based on the timings T1 and T1 ′ (step S27). That is, the timing when one breath starts is T1, and the timing when one breath ends is T1 ′.
As described above, respiratory sound data for one breath is acquired.

  As described above, the breathing sound data processing apparatus 100 acquires the breathing sound data for one breath by the one breathing sound extraction unit 131 and the one breathing sound normalization unit 132 acquires the one breathing sound data. From the breath sound data, calculate the breathing time and breathing sound intensity, calculate the flow rate and / or flow rate for one breath based on the breathing time and breathing sound intensity, and whether the flow rate and / or flow rate has been normalized Further, the secondary noise detection means 133 detects the presence or absence of secondary noise included in the respiratory sound based on the normalized respiratory sound data of the flow rate and / or flow velocity.

  Therefore, the flow rate and / or flow rate for one breath (or the breathing time and breathing sound intensity for one breath) based only on the breathing sound data acquired through the breathing sound signal acquisition means 11 without using a flow meter. .)) Is standardized, and the flow rate and / or flow velocity, or the respiratory sound data in which the respiratory time and respiratory sound intensity for one breath are standardized are analyzed (detection of auxiliary noise, etc.) A respiratory sound data processing apparatus 100 having a small and simple configuration capable of providing accurate diagnosis support can be realized.

  In addition, the description in this Embodiment shows an example of the respiratory sound data processing apparatus and program which concern on this invention, and is not limited to this. The detailed configuration and detailed operation of the respiratory sound data processing apparatus 100 in the present embodiment can be changed as appropriate without departing from the spirit of the present invention.

  For example, when acquiring the breathing sound data for one breath, the one breathing sound extraction unit 131 alternately generates each start timing of exhalation and inspiration every predetermined time interval, and notifies the subject of each start timing. Information (image or sound) for this purpose may be output to the output means 14. In this case, breathing sound data related to expiration and inhalation corresponding to each start timing is acquired as breathing sound data for one breath, so that breathing sound data for one breath can be accurately acquired.

Further, the breathing sound extraction unit 131 calculates a correlation value between the breathing sound data acquired via the first breathing sound signal acquisition unit 111 and the sample respiratory sound data for a preset breath. It may be calculated by cross-correlation or residual sum of squares, etc., and respiratory sound data for one breath may be acquired based on the calculated correlation value. Further, each correlation value is calculated using a plurality of sample respiratory sound data having different breathing time and respiratory sound intensity for one breath, and the breathing sound data of one breath is obtained based on the calculated plurality of correlation values. It may be.
Hereinafter, the correlation value means a cross correlation value, a residual sum of squares, or the like.

Here, the cross-correlation will be described.
First, let X (t) be the waveform of one breath sound, Y (t) be the waveform of one breath in the past, and let X (t) and Y (t + s) be the mutual covariance R _XY (s). Calculate (t and s represent time).

μ _X and μ _Y are time average values of X (t) and Y (t), respectively. The mutual covariance R _XY (s) is a function of only s.

Next, the cross-covariance R _XY (s) is normalized to obtain a cross-correlation function ρ _XY .

R _XX and R _YY are autocovariance functions and are calculated as follows.

In this embodiment, X (t), Y ( t) because it has adapted to the both the same size, it is s = 0 in the cross-correlation function [rho _XY.

  Furthermore, the residual sum of squares means that the difference between the intensity of two breathing sounds at a predetermined time t is squared to calculate the sum of one breath.

  Further, the secondary noise detection means 133 may detect the secondary noise by applying a “hidden Markov process” to one breath.

  Here, to detect the side noise by applying the “hidden Markov process” to one breath, first, FFT processing is performed on the detected breath sound data for one breath for each time segment, and the breath sound is detected. Spectral information for the data is calculated, and then using the spectral information, a spectrogram is calculated for each time segment (eg, early expiration, middle expiration, late expiration, early inspiration, mid inspiration, late inspiration, etc.), Based on the EM algorithm, pre-set learning data, that is, spectrograms for samples (hereinafter referred to as abnormal learning data) for a plurality of respiratory sound data including different types of sub-noise, and a plurality of sub-noises are not included. State transition matrix and appearance probability using sample spectrogram (hereinafter referred to as normal learning data) Sub-noise is identified by classifying the detected respiratory sound data for one breath as the best fit between the case where abnormal learning data is used and the case where normal learning data is used. (Detection). For example, when the abnormal learning data is used more appropriately, the respiratory sound data is classified as respiratory sound data including auxiliary noise related to the abnormal learning data.

  Further, a plurality of past breathing sound data of each subject is stored in advance in the built-in memory (or an external dedicated database not shown), and the sub-noise detection means 133 stores the past breathing sound data and Based on the current breathing sound data of the same subject, the detection of secondary noise is performed on the current respiratory sound data (that is, the presence or absence of secondary noise, the degree of abnormality represented by the secondary noise, etc.). May be. In this case, for example, for the past and current breathing sound data of the same subject, the number of occurrences of intermittent rattle (a kind of sub-noise) per unit time is calculated, and the correlation value related to the number of occurrences is calculated. Or sub-noise is detected by calculating the difference between the two respiratory sound data.

  In addition, the respiratory sound data for samples that does not include secondary noise is stored in advance in the built-in memory (or the dedicated database or the like), and the secondary noise detection means 133 performs the sampling data and the actually acquired respiratory sound. Based on the data, detection of sub-noise (that is, presence / absence of sub-noise, degree of abnormality represented by sub-noise, etc.) may be performed. That is, the difference between the two respiratory sound data is calculated, and sub-noise is detected based on the calculated difference.

  Here, the respiratory sound wave forms represented by the two respiratory sound data to be compared with each other may be displayed side by side on the display screen, and each of the two respiratory sounds may be, for example, stereo “L ( (Left) "and" R (Right) "may be reproduced separately.

  Further, personal information of the subject (for example, information such as height, weight, sex, obesity level, medical history, vital capacity, etc.) is stored in advance in the built-in memory (or the dedicated database, etc.), and one breathing sound normalizing means 132 may correct the flow rate and / or flow rate of one breath based on the personal information. In this case, for example, for a subject with a high degree of obesity, since the respiratory sound collected by the first respiratory sound signal acquisition unit 111 is greatly attenuated due to fat in the body, the flow rate / flow rate calculation unit 1321 The flow rate and / or flow rate of one breath of the subject calculated by is corrected so as to increase by a predetermined amount according to the degree of obesity.

<Modification 1>
The upper respiratory sound data processing apparatus 100 acquires a respiratory sound signal using the respiratory sound signal acquisition unit 11 and acquires respiratory sound data for one breath based on the respiratory sound signal. Not limited to this, as in the respiratory sound data processing apparatus 101 shown in FIG. 4, a plurality of (n; n is a natural number of 2 or more) first respiratory sound signal acquisition means 111 to nth respiratory functions having the same function. Respiratory sound signals are acquired using the sound signal acquisition means 11n (that is, respiratory sounds are acquired from n microphones), and respiratory sound data for one breath is acquired based on the acquired plurality of respiratory sound signals. You may do it.

  Hereinafter, with reference to FIG. 4 and FIG. 5, the respiratory sound data processing for acquiring respiratory sound data for one breath using the first respiratory sound signal acquisition unit 111 to the nth respiratory sound signal acquisition unit 11n will be described. To do. FIG. 4 shows an internal configuration of the respiratory sound data processing apparatus 101 of the first modification, and FIG. 5 shows a flowchart for explaining the respiratory sound data processing of the first modification.

  Here, each component of the breathing sound data processing apparatus 101 shown in FIG. 4 is the same as each component of the breathing sound data processing apparatus 100, and the same reference numerals are given for the sake of brevity and description thereof is omitted.

  In addition, the first respiratory sound signal acquisition unit 111 to the nth respiratory sound signal acquisition unit 11n are arranged in a predetermined part of the body of the subject. For example, when the first respiratory sound signal acquisition unit 111 and the second respiratory sound signal acquisition unit 112 are used, the first respiratory sound signal acquisition unit 111 is disposed near the bronchus and the second respiratory sound signal acquisition unit 112 is used. The acquisition means 112 may be arranged near the lower lobe of the lung, and when the first respiratory sound signal acquisition means 111 to the fifth respiratory sound signal acquisition means 115 are used, the first respiratory sound signal The acquisition means 111 is arranged near the bronchi, the second respiratory sound signal acquisition means 112 and the third respiratory sound signal acquisition means 113 are arranged near the left and right upper lobes of the lung, respectively, and the fourth respiratory sound signal acquisition means 114 The fifth respiratory sound signal acquisition means 115 may be mounted near the left and right lower lobes of the lung.

  Next, with reference to the flowchart of FIG. 5, the breathing sound data processing in the first modification will be described.

  The data processing means 13 executes the respiratory sound data processing shown in the flowchart of FIG.

  First, the breathing sound extraction means 131 acquires a plurality of breathing sound data via the first breathing sound signal acquisition means 111 to the nth breathing sound signal acquisition means 11n and the A / D conversion means 12 (step S30). ), Whether the magnitude of the breathing sound represented by the breathing sound data acquired via the first breathing sound signal acquisition unit 111 exceeds a preset threshold Th1 (that is, whether one breath has started) Is monitored (step S31).

  When the intensity of the breathing sound exceeds the threshold Th1 (step S31; Yes), the breathing sound extraction unit 131 detects the timing T1 at this time (that is, at the start of one breathing) (step S32). Data indicating the detected timing T1 is temporarily stored in the internal memory.

  After step S32, the breathing sound extraction means 131 has a plurality of breathing sound levels represented by the breathing sound data acquired from each of the second breathing sound signal acquisition means 112 to the nth breathing sound signal acquisition means 11n. Among them, whether or not a predetermined number of breath sounds (for example, two or more may be set freely) exceeds the corresponding threshold values Th2 to Thn (that is, during one breath). Whether or not) is monitored (step S33).

  When the magnitude of a predetermined number of breathing sounds out of the plurality of breathing sounds exceeds the corresponding threshold values Th2 to Thn (step S33; Yes), the breathing sound extraction means 131 detects the timing T2 at this time. (Step S34). Data indicating the detected timing T2 is temporarily stored in the built-in memory.

  After step S34, the breathing sound extraction unit 131 determines whether or not the magnitude of the breathing sound represented by the breathing sound data acquired via the first breathing sound signal acquisition unit 111 has exceeded the threshold Th1 again ( That is, whether or not one breath has been completed is monitored (step S35), and when the volume of the breathing sound exceeds the threshold value Th1 again (step S35; Yes), at this time (that is, at the end of one breath) The timing T1 ′ is detected (step S36).

After step S36, the breathing sound extraction unit 131 identifies one breath based on the timings T1 and T1 ′ (step S37). That is, the timing when one breath starts is T1, and the timing when one breath ends is T1 ′.
As described above, respiratory sound data for one breath is acquired.

  As described above, since one breath is specified based on a plurality of breathing sound data obtained from a plurality of locations of the subject's body, compared to the case where only one breathing sound data is used, more accurately and more reliably for one breathing. Data can be extracted.

<Modification 2>
In addition, the respiratory sound data processing devices 100 and 101 have standardized the breathing time and respiratory sound intensity for one breath by standardizing the flow rate and / or flow velocity. Like the breathing sound data processing apparatus 102 shown, the breathing time and breathing sound intensity for one breath may be normalized directly and automatically.

  Hereinafter, the respiratory sound data processing apparatus 102 according to the third modification will be described in detail with reference to FIGS. 6 and 7. FIG. 6 shows the internal configuration of the respiratory sound data processing apparatus 101 of the second modification, and FIG. 7 shows a flowchart for explaining the respiratory sound intensity normalization process of the second modification.

  Here, among the constituent parts of the respiratory sound data processing apparatus 102 shown in FIG. 6, the parts other than the data processing means 13a are the same as the constituent parts of the respiratory sound data processing apparatuses 100 and 101, and are the same for simplification. The reference numerals are attached and the description is omitted. Further, in the data processing means 13a, each component other than the one breathing sound normalization means 132a is the same as each component provided in the data processing means 13 of the breathing sound data processing devices 100 and 101, and for the sake of simplification. The same reference numerals are given and the description is omitted.

  As shown in FIG. 6, the breathing sound normalization means 132a includes a breathing sound intensity normalization means 1321a and a breathing time normalization means 1321b, and the breathing sound intensity normalization means 1321a is the breathing sound shown in the flowchart of FIG. Data processing is performed to normalize the respiratory sound intensity from the breathing sound data for one breath input from the breathing sound extraction unit 131, and the breathing time normalization unit 1321b is a dynamic time warping (DTW). The breathing time is normalized from the breathing sound data for one breath inputted from the one breathing sound extraction means 131 using a time normalization method called “respiration time”.

  Here, the breathing sound normalization means 132a may be hardware as in the breathing sound extraction means 131 and the sub noise detection means 133, or a program executed by the data processing means 13a has. It may represent a function. Here, it is assumed that the program is a program for performing the respiratory sound data processing shown in the flowchart of FIG. 7, and is stored in advance in a built-in memory (not shown) of the data processing means 13a.

  The standardization of the breathing time using the DTW is time information (for example, the first period of expiratory period (further, the first period, the middle period, and the second period). Similarly, multiple time segments such as mid-expiration, late exhalation, early inspiration, mid-inspiration, late inspiration, etc.) for each time segment to match the standardized sample time information set in advance. This is shaping the waveform of respiratory sound data.

Next, the breathing sound data processing in the second modification will be described with reference to the flowchart in FIG.
The breathing sound intensity normalization means 1321a uses the low-pass respiratory sound data corresponding to a low frequency equal to or lower than a predetermined reference frequency using the LPF (Low Pass Filter) (not shown) as input breathing sound data for one breath. Then, the high frequency breathing sound data corresponding to the other high frequency is classified (step S40). Here, the LPF is a logic circuit or a program.

  Next, the respiratory sound intensity normalization means 1321a determines the intensity of the respiratory sound related to the separated low frequency respiratory sound data so as to match the intensity of the low frequency component of the preset standardized respiratory sound data. Adjustment (normalization) is performed (step S41).

  Note that an LPF that separates the respiratory sound signal acquired from the respiratory sound signal acquisition unit 11 into a low-frequency component and a high-frequency component may be provided after the respiratory sound signal acquisition unit 11.

  In this case, since the breathing sound intensity normalization means 1321a does not need to separate the breathing sound data converted into digital data into low-frequency breathing sound data and high-frequency breathing sound data, data processing in the breathing sound intensity normalization means 1321a Time can be shortened.

Thereafter, the respiratory sound intensity normalization means 1321a synthesizes the high frequency component and the normalized low frequency component (step S42).
As described above, the intensity of the respiratory sound data is normalized.

  Although the respiratory sound data processing devices 100, 101, and 102 have been described above, these may be realized individually or in combination with each other.

  Further, the program for performing the breathing sound data processing and the breathing sound data may be stored in advance in various recording media such as a CD (Compact Disk) or acquired from an external device via a network. You may be able to do it. In this case, each of the respiratory sound data processing devices 100, 101, and 102 transmits / receives data by connecting to a dedicated driver or a network for freely reading / writing data to / from the recording medium. For communication (none is shown).

It is a block diagram which shows the internal structure of the respiratory sound data processing apparatus to which this invention is applied. It is a flowchart explaining the respiratory sound data processing by the respiratory sound data processing apparatus to which this invention is applied. It is a flowchart for demonstrating the respiration sound data processing by the respiration sound data processing apparatus to which this invention is applied. It is a block diagram which shows the internal structure of the other respiratory sound data processing apparatus to which this invention is applied. It is a flowchart for demonstrating the respiration sound data processing by the respiration sound data processing apparatus shown in FIG. It is a block diagram which shows the internal structure of the other respiratory sound data processing apparatus to which this invention is applied. It is a flowchart explaining the respiratory sound data processing by the respiratory sound data processing apparatus shown in FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Respiratory sound signal acquisition means 111-11n 1st-nth respiratory sound signal acquisition means 12 A / D conversion means 13, 13a Data processing means 14 Output means 100-102 Respiratory sound data processing device 131 One respiratory sound extraction means 132 132a One breathing sound normalization means 133 Sub noise detection means 1321 Flow rate / flow velocity calculation means 1321a Breathing sound intensity normalization means 1321b Breathing time normalization means

Claims (48)

A respiratory sound data processing apparatus for analyzing and processing respiratory sound data representing respiratory sounds,
One breath sound extraction means for extracting data for one breath from the breath sound data;
A breathing sound normalization means for calculating a breathing time and breathing sound intensity from the breathing sound data for one breath and normalizing the breathing sound data for the breathing sound based on the calculated breathing time and breathing sound intensity When,
A respiratory sound data processing apparatus comprising: Breathing sound signal acquisition means for acquiring a breathing sound signal representing the breathing sound;
A / D conversion means for digitizing the respiratory sound signal acquired by the respiratory sound signal acquisition means into respiratory sound data;
Further comprising
The breathing sound extraction means includes
2. The respiratory sound data processing apparatus according to claim 1, wherein data for one breath is extracted from the respiratory sound data digitized by the A / D conversion means. The breathing sound extraction means includes
When two consecutive timings when the magnitude of the respiratory sound related to the respiratory sound data acquired via the respiratory sound signal acquisition unit exceeds a preset threshold are detected, the two detected timings are used. The respiratory sound data processing apparatus according to claim 1 or 2, wherein one breath is specified and data for one breath is extracted from the respiratory sound data. The breathing sound extraction means includes
A correlation value between the respiratory sound data and preset sample respiratory sound data for one breath is calculated, and data for one breath is extracted from the respiratory sound data based on the calculated correlation value. The respiratory sound data processing device according to claim 1 or 2. The breathing sound extraction means includes
The correlation value is calculated using a plurality of preset sample respiratory sound data having different breathing time and / or respiratory sound intensity for one breath, and data for one breath is calculated based on the calculated plurality of correlation values. The respiratory sound data processing device according to claim 4, wherein: Comprising first to nth (n is a natural number of 2 or more) respiratory sound signal acquisition means for acquiring a respiratory sound signal representing a respiratory sound;
The A / D conversion means includes
2. The respiratory sound data processing apparatus according to claim 1, wherein each of the plurality of respiratory sound signals acquired by the first to nth respiratory sound signal acquisition means is digitized into respiratory sound data. The breathing sound extraction means includes
The timing when the magnitude of the breathing sound related to the breathing sound data acquired via the first breathing sound signal acquisition unit exceeds a predetermined threshold set in advance according to the first breathing sound signal acquisition unit. After the detection, the second to nth respiratory sound signal acquisition means has a magnitude of the respiratory sound related to the respiratory sound data acquired through at least one of the second to nth respiratory sound signal acquisition means. Each of which exceeds a predetermined threshold value set in advance according to each of the first and second respiratory sounds, and the magnitude of the respiratory sound related to the respiratory sound data acquired through the first respiratory sound signal acquisition means is the first respiratory sound. The respiratory sound data according to claim 6, wherein when the timing when the signal acquisition means exceeds the threshold again is detected, data for one breath is extracted from the respiratory sound data using the two timings. Processing equipment. The breathing sound extraction means includes
Each start timing of exhalation and inspiration is alternately generated and notified every predetermined time interval, and breathing sound data acquired according to the start timing of expiration is used as breathing sound data corresponding to expiration, and the start timing of inspiration The respiratory sound data processing apparatus according to claim 1, wherein the respiratory sound data for one breath is acquired as the respiratory sound data corresponding to inspiration as the respiratory sound data acquired accordingly. The breathing sound normalization means includes:
Flow rate / flow velocity calculation means for calculating a flow rate and / or flow velocity for one breath based on the breathing time and breathing sound intensity for one breath is further provided, and one breathing sound is calculated based on the calculated flow rate and / or flow velocity. The respiratory sound data processing apparatus according to any one of claims 1 to 8, wherein the respiratory sound data for a minute is normalized. The breathing sound normalization means includes:
The respiratory sound data processing device according to claim 9, wherein a flow rate for one breath is calculated by multiplying a breathing sound intensity by a breathing time for one breath. The breathing sound normalization means includes:
The respiratory sound data processing apparatus according to claim 9 or 10, wherein the respiratory sound intensity for one breath is divided by the respiratory time to calculate a flow velocity for the one breath. The breathing sound normalization means includes:
The respiratory sound data processing apparatus according to any one of claims 9 to 11, wherein the flow rate and / or flow velocity is corrected using personal information for each subject. The breathing sound normalization means includes:
The respiratory sound data processing according to claim 12, wherein the flow rate and / or flow rate is corrected using personal information including at least one of height, weight, sex, obesity level, medical history, and vital capacity of a subject. apparatus. The breathing sound normalization means includes:
It is determined whether or not the flow rate and / or flow velocity calculated by the flow rate / velocity calculation means matches a preset standard and flow rate and / or flow velocity. The respiratory sound data processing device according to any one of claims 9 to 13, wherein the breathing sound data processing device is notified. The breathing sound normalization means includes:
Breathing sound intensity normalization means for normalizing the breathing sound intensity related to breathing sound data for one breath;
Breathing time normalization means for normalizing the breathing time related to breathing sound data for one breath;
With
The respiration sound data according to any one of claims 1 to 8, wherein the respiration sound data for one breath is normalized using the respiration sound intensity normalization means and the respiration time normalization means. Sound data processing device. The breathing sound intensity normalization means is
The respiratory sound data for one breath is separated into low-frequency respiratory sound data whose respiratory sound frequency is equal to or lower than a predetermined reference frequency and high-frequency respiratory sound data exceeding the reference frequency, and the respiratory sound intensity of the low-frequency respiratory sound data is determined. 16. The normalization of the low-frequency breathing sound data after the normalization and the high-frequency breathing sound data to normalize the breathing sound intensity related to the breathing sound data for the one breath. The respiratory sound data processing device described. The breathing sound intensity normalization means is
Among the low-frequency respiratory sound data whose respiratory sound frequency is preliminarily separated from the respiratory sound data for one breath and the high-frequency respiratory sound data whose frequency exceeds the reference frequency, the respiratory sound of the low-frequency respiratory sound data 16. The intensities are normalized, and the normalized low-frequency breathing sound data is combined with the high-frequency breathing sound data to normalize the breathing sound intensity related to the breathing sound data for the one breath. 2. A respiratory sound data processing device according to 1. The breathing time normalization means includes
The respiratory sound data processing apparatus according to any one of claims 15 to 17, wherein the respiratory sound data for one breath is normalized using a dynamic time warping method.   The sub-noise detecting means for detecting the sub-noise contained in the breathing sound from the standardized breathing sound data for one breath is further provided. Respiratory sound data processing device. The sub-noise detection means includes
Spectral information related to the breathing sound data is obtained by performing FFT (Fast Fourie Transformation) processing on the standardized breathing sound data for one breath, and applying a hidden Markov process to the spectral information, sub-noise The respiratory sound data processing device according to claim 19, wherein the detection is performed. The sub-noise detection means includes
Calculating a correlation value between two normalized breathing sound data for one breath acquired from the same subject at two different times, and detecting sub-noise based on the calculated correlation value The respiratory sound data processing device according to claim 19. The sub-noise detection means includes
A correlation value between the standardized breathing sound data for one breath and the sampled breathing sound data not including the secondary noise is calculated, and the secondary noise is detected based on the calculated correlation value. The respiratory sound data processing apparatus according to claim 19.   The respiratory sound data processing apparatus according to any one of claims 4, 5, and 9 to 22, wherein the correlation value is a cross-correlation value or a residual sum of squares. A computer for analyzing and processing respiratory sound data representing respiratory sounds,
A function of extracting data for one breath from the breathing sound data;
A function for calculating a breathing time and a breathing sound intensity from the breathing sound data for the one breath, and normalizing the breathing sound data for the one breathing sound based on the calculated breathing time and the breathing sound intensity;
A program to realize In the computer,
When two consecutive timings when the magnitude of the respiratory sound related to the respiratory sound data exceeds a preset threshold are detected, one respiratory is identified using the two detected timings, and the respiratory sound data The program according to claim 24, further realizing a function of extracting data for one breath from the data. In the computer,
A function of calculating a correlation value between the respiratory sound data and preset sample respiratory sound data for one breath, and further extracting data for one breath from the respiratory sound data based on the calculated correlation value The program according to claim 24 for realizing. In the computer,
The correlation value is calculated using a plurality of preset sample respiratory sound data having different breathing time and / or respiratory sound intensity for one breath, and data for one breath is calculated based on the calculated plurality of correlation values. 25. The program according to claim 24, for further realizing a function of extracting. In the computer,
Among the respiratory sound data acquired via the first to nth respiratory sound signal acquisition means, the magnitude of the respiratory sound related to the respiratory sound data acquired via the first respiratory sound signal acquisition means Obtained via at least one of the second to nth respiratory sound signal acquisition means after detecting a timing when a predetermined threshold value preset in accordance with the first respiratory sound signal acquisition means is exceeded. The magnitude of the breathing sound related to the breathing sound data exceeds each of the predetermined threshold values set in advance according to each of the second to nth breathing sound signal acquisition means, and then the first breathing sound signal acquisition means When the timing at which the magnitude of the breathing sound related to the breathing sound data acquired via the above again exceeds the threshold value of the first breathing sound signal acquisition means is detected, the breathing is performed using the two detected timings. One breath data from sound data Program according to claim 24 for further realizing the function of extracting. In the computer,
Each start timing of exhalation and inspiration is alternately generated and notified every predetermined time interval, and breathing sound data acquired according to the start timing of expiration is used as breathing sound data corresponding to expiration, and the start timing of inspiration 25. The program according to claim 24, for further realizing a function of acquiring respiratory sound data for one breath as respiratory sound data corresponding to inspiration using the respiratory sound data acquired in response. In the computer,
Flow rate / flow velocity calculation means for calculating a flow rate and / or flow velocity for one breath based on the breathing time and breathing sound intensity for one breath is further provided, and one breathing sound is calculated based on the calculated flow rate and / or flow velocity. 30. The program according to any one of claims 24 to 29, for further realizing a function of normalizing minute respiratory sound data. In the computer,
The program according to claim 30, further realizing a function of calculating a flow rate for one breath by multiplying a breathing sound intensity by a breathing time for one breath. In the computer,
32. The program according to claim 30 or 31 for further realizing a function of dividing a respiratory sound intensity for one breath by a breathing time and calculating a flow velocity for the one breath. In the computer,
The program as described in any one of Claims 30-32 for further implement | achieving the function which correct | amends the said flow volume and / or flow velocity using the personal information for every test subject. In the computer,
34. The program according to claim 33, further realizing a function of correcting the flow rate and / or flow rate using personal information including at least one of a subject's height, weight, sex, obesity level, medical history, and vital capacity. In the computer,
It is determined whether or not the flow rate and / or flow velocity calculated by the flow rate / velocity calculation means matches a preset standard and flow rate and / or flow velocity. The program as described in any one of Claims 30-34 for further implement | achieving the function which alert | reports. In the computer,
A function to normalize the breathing sound intensity associated with breathing sound data for one breath;
A function for normalizing the breathing time associated with breathing sound data for one breath;
The program as described in any one of Claims 24-29 for further implement | achieving. In the computer,
Respiratory sound intensity of the low-frequency respiratory sound data is standardized by classifying the respiratory sound data for one breath into low-frequency respiratory sound data whose respiratory sound frequency is equal to or lower than a predetermined reference frequency and high-frequency respiratory sound data exceeding the reference frequency. 37. The method of further realizing a function of normalizing the respiratory sound intensity related to the respiratory sound data for one breath by synthesizing the normalized low frequency respiratory sound data and the high frequency respiratory sound data. The program described in. In the computer,
The respiratory sound intensity of the low-frequency respiratory sound data among the low-frequency respiratory sound data whose respiratory sound frequency is preliminarily separated from the respiratory sound data for one breath and the high-frequency respiratory sound data that exceeds the reference frequency And further normalizing the respiratory sound intensity of the respiratory sound data for one breath by combining the normalized low-frequency respiratory sound data with the high-frequency respiratory sound data. 36. The program according to 36. In the computer,
The program according to any one of claims 36 to 38, for further realizing a function of normalizing a breathing time using a dynamic time warping method for breathing sound data for one breath. In the computer,
The program according to any one of claims 24 to 39, for further realizing a function of detecting sub-noise included in the breathing sound from the normalized breathing sound data for one breath. In the computer,
A function of performing FFT processing on the normalized breathing sound data for one breath to obtain spectrum information related to the breathing sound data, and detecting sub-noise by applying a hidden Markov process to the spectrum information 41. The program according to claim 40 for further realizing the above. In the computer,
A function to calculate the correlation value between two standardized breathing sound data for one breath acquired from the same subject at two different times and to detect sub-noise based on the calculated correlation value is further realized. 41. The program according to claim 40 for causing the program to occur. In the computer,
A correlation value between the standardized breathing sound data for one breath and the sampled breathing sound data not including the secondary noise is calculated, and the secondary noise is detected based on the calculated correlation value. 41. The program according to claim 40 for further realizing a function.   44. The program according to any one of claims 26, 27, and 30 to 43, wherein the correlation value is a cross-correlation value or a residual sum of squares.   23. Respiratory sound data according to claim 21 or 22, further comprising output means for displaying each of the respiratory sound waveforms represented by the two respiratory sound data and / or reproducing the two respiratory sounds. Processing equipment. The output means includes
46. The respiratory sound data processing apparatus according to claim 45, wherein the respiratory sound wave forms represented by the two respiratory sound data are displayed side by side. In the computer,
44. The program according to claim 42 or 43, for further realizing a function of displaying each of the respiratory sound waveforms represented by the two respiratory sound data and / or reproducing the two respiratory sounds. In the computer,
48. The program according to claim 47, further realizing a function of displaying each respiratory sound waveform represented by the two respiratory sound data side by side.

Images