JP2000093399A - Phonocardiograph - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a phonocardiograph which achieves an accurate diagnosis of heart diseases. SOLUTION: A cardiac sound waveform in a specified section containing a II sound generated per beat is extracted by a waveform extraction means 42 as extracted waveform WEn from a cadiac sound signal SH and phases of the waveform WEn are put together by a phase matching means 58 to match mutually. The extracted waveform WEn thus having the phases matched is added by an addition means 60 (SA9) to let a normal cardiac sound overlap a cardiac noise generated per beat to equalize an external noise. A typical extraction waveform determining means 62 determines a typical extraction waveform WR based on the waveform obtained by addition by the addition means 60 thereby achieving an accurate diagnosis of heart diseases using the typical extraction waveform WR.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electrocardiograph for detecting a heart sound generated from the heart of a living body and measuring a heart sound chart for diagnosing a heart disease.

[0002]

2. Description of the Related Art In order to examine a heart disease, a doctor performs direct auscultation by an auscultation method. In the diagnosis using the auscultation method, since the volume and tone of heart sounds are different due to individual differences between doctors who auscultate, objectivity and quantitativeness are lacking, and heart sounds cannot be recorded.

[0003] Therefore, a heart sound sensor attached to the body surface is provided as an alternative to the auscultation method or a means for supplementing the auscultation method, and a heart sound detected from the heart sound sensor is amplified and amplified.
Measuring the phonocardiogram using a recording phonograph is performed. The phonocardiogram measured by the phonograph can be obtained as objective information without individual differences, and can be recorded.The phonocardiogram of a patient with heart disease can be obtained in addition to the normal heart sound. Since a heart murmur generated due to the disease is included, a heart disease can be diagnosed.

[0004]

However, since the above-mentioned conventional heart sound meter merely amplifies and records the heart sound detected from the heart sound sensor, the measured heart sound chart shows a normal heart sound and a heart disease. In addition to the cardiac murmur generated due to the above, there is included external noise, that is, noise other than the sound from the heart, for example, respiratory noise accompanying breathing, noise from outside the body such as footsteps, opening and closing of doors, and the like. For this reason, it has sometimes been difficult to accurately diagnose a heart disease from a heart sound chart.

[0005] The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a phonograph which can accurately diagnose a heart disease.

[0006]

Means for Solving the Problems The present inventor has made various studies on the background described above, and the external murmur does not occur every beat unlike normal heart sounds and murmurs. Since it occurs at a plurality of beat cycles of about 4 to 5 beats like respiratory noise or happens accidentally like noise from outside the body, by adding a heart sound waveform, We found that the influence could be reduced.

That is, the gist of the present invention for achieving the above object is to provide a heart sound sensor which detects a heart sound generated from the heart and outputs a heart sound signal representing the heart sound, and a heart sound sensor represented by the heart sound signal. A heart sound meter for measuring a waveform, wherein (a) from the heart sound signal, a waveform extracting means for extracting a heart sound waveform of a predetermined section generated in each cycle as an extracted waveform, and (b) a waveform extracting means extracted by the waveform extracting means. Phase matching means for adjusting the phases of the extracted waveforms so that the extracted waveforms coincide with each other; (c) adding means for adding the extracted waveforms whose phases have been matched by the phase matching means; and (d) the adding means. A representative extracted waveform determining means for determining a representative extracted waveform based on the calculated extracted waveform.

[0008]

According to this configuration, the waveform extracting means extracts a heart sound waveform of a predetermined section generated every cycle from the heart sound signal as an extracted waveform, and the phase matching means makes the phases of the extracted waveforms coincide with each other. By adding the extracted waveforms whose phases have been adjusted by the adding means, normal heart sounds and heart noises generated for each beat are superimposed, and the external noise is averaged. Since the representative extracted waveform determining means determines the representative extracted waveform based on the waveform obtained by the addition in the adding means, it is possible to accurately diagnose a heart disease using the representative extracted waveform. it can.

[0009]

In another preferred embodiment of the present invention, the heart sound meter preferably comprises:
The apparatus further includes external noise calculation means for calculating an external noise waveform from a difference between the extracted waveform extracted by the waveform extraction means and the representative extracted waveform. According to this configuration, the external noise calculating means calculates the external noise waveform indicating the magnitude of the external noise. Therefore, the external noise waveform can be analyzed or analyzed for external noise such as elucidation of the cause of the external noise. Can be used.

Preferably, the electrocardiograph further includes a representative extracted waveform display means for displaying a representative extracted waveform determined by the representative extracted waveform determining means on a display. With this configuration, the representative extracted waveform is displayed on the display by the representative extracted waveform display means, so that a heart disease can be diagnosed from the displayed representative extracted waveform.

Preferably, in the two-dimensional coordinate system comprising a time axis and an amplitude axis, the heart sound meter is determined by the extracted waveform extracted by the waveform extracting means and by the representative extracted waveform determining means. And comparative display means for simultaneously displaying the representative extracted waveforms so that they can be compared with each other. In this way, the comparison display means displays the extracted waveform extracted by the waveform extraction means and the representative extracted waveform on the two-dimensional coordinate system including the time axis and the amplitude axis so that they can be simultaneously compared. The external noise can be recognized from the difference between the extracted waveform and the representative extracted waveform.

Preferably, the heart sound meter includes external noise display means for displaying an external noise waveform calculated by the external noise calculation means on a display. With this configuration, the external noise display means displays the external noise waveform calculated by the external noise calculation means on the display device.
There is an advantage that the magnitude of the external noise can be easily recognized.

Preferably, the heart sound meter includes an amplitude ratio calculating unit that calculates an amplitude ratio between a maximum amplitude and a minimum amplitude of the extracted waveform extracted by the waveform extracting unit, and the amplitude ratio calculating unit calculates the amplitude ratio. Average value calculating means for calculating the average value of the calculated amplitude ratio, standard deviation calculating means for calculating the standard deviation of the amplitude ratio calculated by the amplitude ratio calculating means, and the amplitude ratio calculated by the amplitude ratio calculating means. Among the extracted waveforms, the deviation from the average calculated by the average calculation means is within a predetermined experimentally predetermined range of the standard deviation calculated by the standard deviation calculation means. Reference extraction waveform determination means for determining an extraction waveform having the maximum amplitude ratio as a reference extraction waveform, wherein the phase matching means includes a base extraction unit that determines the reference extraction waveform determined by the reference extraction waveform determination means. The extraction waveform is intended to match the phase of the extracted waveform to match each other based on the extracted waveform. With this arrangement, the amplitude ratio calculating means calculates the amplitude ratio of the extracted waveform extracted by the waveform extracting means, the average value calculating means calculates the average value of the amplitude ratio, and the standard deviation calculating means calculates the amplitude ratio. The standard deviation of the ratio is calculated, and in the reference extraction waveform determining means, among the extracted waveforms, the amplitude ratio is a deviation from the average value within a predetermined experimentally predetermined range of the standard deviation. An extracted waveform having the maximum amplitude ratio is determined as a reference extracted waveform. That is, from the extracted waveforms excluding the abnormally large amplitude ratio such as spike noise, the extracted waveform with the least external noise is determined as the reference extracted waveform. The phase matching means adjusts the phases of the extracted waveforms based on the reference extracted waveforms so that the extracted waveforms match each other, so that the phases of the extracted waveforms can be matched with each other with high accuracy. Therefore, the extracted waveforms are added by the adding means, and a representative extracted waveform determined based on the waveform obtained by the addition by the representative extracted waveform determining means is a heart sound and a heart noise generated for each beat. Has the advantage of being further emphasized with respect to the external noise.

[0014] Preferably, the electrocardiograph includes reference extraction waveform updating means for updating a reference extraction waveform based on the extraction waveform added by the addition means, and the phase matching means includes a reference extraction waveform. The phase of the reference extracted waveform updated by the updating unit and the phase of the extracted waveform that has not been added by the adding unit among the extracted waveforms extracted by the waveform extracting unit match each other, and the adding unit is The reference extraction waveform and the extraction waveform whose phases are mutually matched by the phase matching means are added. By doing so, the phase matching means
The phase of the reference extracted waveform updated by the reference extracted waveform updating means and the phase of the extracted waveform before being added by the adding means are matched with each other, and the adding means matches the phases with each other by the phase matching means. The extracted reference waveform and the extracted waveform are added, and the reference extraction waveform updating means updates the reference extraction waveform based on the extraction waveform added by the adding means. Therefore, for each extracted waveform, the phase of the extracted waveform and the phase of the sequentially extracted reference extracted waveform are matched with each other, so that when the phases of the extracted waveforms are matched based on one reference extracted waveform, Since the phase matching accuracy is improved in comparison, there is an advantage that a representative extracted waveform in which heart sounds and heart murmurs are further emphasized with respect to external noise is obtained.

Preferably, the heart sound meter includes a time-frequency analyzing means for analyzing the representative extracted waveform with respect to time and frequency. With this configuration, the time-frequency analysis unit analyzes the representative extracted waveform with respect to time and frequency, so that a normal heart sound and heart murmur included in the representative extracted waveform can be separated, and the heart murmur can be separated. Since the time of occurrence can be analyzed, there is an advantage that diagnosis based on a representative extracted waveform can be easily performed. For example, it is possible to easily know the presence of a heart murmur and to specify a site where the heart murmur occurs.

Preferably, the heart sound meter includes external noise analysis means for frequency-analyzing the external noise waveform. With this configuration, the external noise waveform is frequency-analyzed by the external noise analysis unit, and a plurality of signal components such as the respiratory noise and the extraneous noise included in the external noise waveform are separated. Therefore, it is possible to diagnose a respiratory disease using the analysis spectrum of the external noise waveform subjected to the frequency analysis, and it is easy to clarify the cause of the noise from outside the living body. It is easier to eliminate the cause.

[0017]

Preferred embodiments of the present invention will be described below in detail with reference to the accompanying drawings. FIG. 1 is a block diagram illustrating a configuration of a heart sound meter 10 to which the present invention is applied.

In FIG. 1, the microphone 12 functions as a heart sound sensor, and is fixed on the chest of the subject 14 with an adhesive tape (not shown) or the like. Then, in a piezoelectric element provided inside the microphone 12 (not shown), a heart sound or the like generated from the heart of the subject 14 is converted into an electric signal, that is, a heart sound signal SH.
The heart sound signal amplifier 16 is provided with four types of filters (not shown) that attenuate high-energy bass components in order to well record the high-tone components of the heart sound, and the heart sound signal SH output from the microphone 12 is amplified. After being waved, it is supplied to the electronic control unit 20 via the A / D converter 18.

One set of electrodes 22 is attached to a predetermined part of the living body in order to output an electrocardiographic signal SE indicating an action potential of the myocardium. In the present embodiment, in order to measure an electrocardiogram by lead II, the subject 14 is attached to the right hand and the left foot of the subject 14, respectively. Electrocardiographic signal SE output from electrode 22
Is amplified by the electrocardiographic signal amplifier 24 and then supplied to the electronic control unit 20 via the A / D converter 26. The electronic control unit 20 is supplied with a start signal SS from a push button 28 and a pulse signal SP with a predetermined frequency from a clock signal source 29.

The electronic control unit 20 includes a CPU 30, R
The microcomputer 30 includes a so-called microcomputer having an OM 32, a RAM 34, an I / O port (not shown), and the like. The CPU 30 executes signal processing using a storage function of the RAM 34 according to a program stored in the ROM 32 in advance. Thus, a heart sound waveform represented by the heart sound signal SH, that is, a heart sound diagram and an electrocardiogram signal S are displayed on the display screen of the display unit 36.
An electrocardiographic lead waveform represented by E, that is, an electrocardiogram is displayed, and the heart sound waveform is analyzed, and the result is displayed.

FIG. 2 is a functional block diagram for explaining a main part of the control function of the electronic control unit 20 in the phonograph 10. As shown in FIG. In FIG. 2, a waveform storage means 40 stores a heart sound signal SH and an electrocardiogram signal SE together with an elapsed time t counted from the start of measurement counted based on the pulse signal SP from the clock signal source 29 in a predetermined storage (not shown) in the RAM 34. Store in the area.

The waveform extracting means 42 stores a predetermined period or a predetermined number of beats (for example, 3 beats) stored in the waveform storing means 40.
From the heart sound signal SH of 0 second or 30 beats), a heart sound waveform of a predetermined section generated in each cycle is extracted, that is, an extraction waveform WE is extracted. FIG. 3 is an example of a heart sound chart and an electrocardiogram represented by the heart sound signal SH and the electrocardiogram signal SE stored in the RAM 34 by the waveform storage means 40.
3A shows an electrocardiogram, FIG. 3B shows a normal electrocardiogram, and FIG.
(C) shows a heart sound chart when there is aortic stenosis, and FIG.
(D) each characteristically shows a heart sound chart in the case of left atrioventricular valve insufficiency. As shown in FIG. 3B, a normal heart sound diagram includes sounds I to IV. Therefore, the waveform extracting means 42 determines, for example, the rising of the II sound from the heart sound signal SH stored in the waveform storing means 40, and extracts a heart sound waveform in a predetermined section from the rising point of the II sound as the extracted waveform WE. Extract as Alternatively, an electrocardiogram measured simultaneously with the electrocardiogram as a reference lead is used, and a heart sound signal in a predetermined section after a predetermined time has elapsed from the detection of a predetermined portion (for example, an R wave) which is periodically detected in the electrocardiogram. The heart sound waveform represented by SH is extracted as an extraction waveform WE.

The amplitude ratio calculating means 44 calculates the extracted waveform W shown in FIG.
As shown in an example of E, it determines the maximum amplitude D _s and the minimum amplitude D _n for each extraction waveform WE extracted by the waveform extraction unit 42, the maximum amplitude D _s and the minimum amplitude D _n
Is calculated (D _S / D _n ).

The average value calculating means 46 includes an amplitude ratio calculating means 4
Average value D of the amplitude ratio calculated in 4 (D _S / D _n)
AV is calculated, and the standard deviation calculating means 48 calculates a standard deviation s of the amplitude ratio (D _S / D _n ) calculated by the amplitude ratio calculating means 44.

The reference extraction waveform determining means 54 calculates the amplitude ratio (D _S / D _n ) calculated by the amplitude ratio calculating means 44,
The extracted waveform WE whose deviation from the average value D _AV calculated by the average value calculating means 46 is within a predetermined experimentally predetermined range of the standard deviation s calculated by the standard deviation calculating means 48. , to determine the extracted waveform WE whose amplitude ratio (D _S / D _n) is the maximum as a reference extraction waveform W _ST. That is, the extracted waveform WE in which the deviation of the amplitude ratio (D _S / D _n ) from the average value D _AV is not within a predetermined experimentally predetermined multiple (for example, twice) of the standard deviation s is excluded. , The amplitude ratio of spike noise and the like (D _S / D _n )
The extracted waveform WE having the largest amplitude ratio (D _S / D _n ) among the extracted waveforms WE excluding the extracted waveform WE having an abnormally large extracted waveform WE such as spike noise is excluded.
_Is determined as the reference extraction waveform _WST . Therefore, the reference extraction waveform W _ST determined in this manner is used as the waveform extraction means 4.
2 is the extracted waveform WE having the least external noise among the normal extracted waveforms WE extracted in Step 2.

The permutation determining means 56 is a
The extracted waveform WE is used as a phase matching
The order of processing in the stage 58 and the adding means 60 is
The amplitude ratio (D) calculated by the width ratio calculating means 44_S/ D
_n), The average value D calculated by the average value calculating means 46
_AV, And the standard calculated by the standard deviation calculating means 48.
Determined based on the quasi-deviation s. That is, the amplitude ratio (D_S
/ D_n) Is used in the reference extraction waveform determination means 54.
Range (ie the average value D_AVIs the standard deviation s
Range of a predetermined multiple of experimentally determined in advance) is defined as a first range.
Then, a range outside the first range is set as a second range, and phase adjustment is performed.
The extracted waveform WE is processed by the means 58 and the adding means 60.
In the first range, the reference extraction waveform WE_ST
Except for the amplitude ratio (D_S/ D_n) From larger to smaller
Then, in the second range, the amplitude ratio (D_S/ D_n) Large
The amplitude ratio (D_S/ D_n) To the minimum value
You. Therefore, the phase matching means 58 and the adding means 60
Is relatively the reference extraction waveform WE _STExtracted waveform W similar to
Since the processing is performed in order from E,
Noise can be prevented and the effects of external noise can be
You.

The phase matching means 58 includes a waveform extracting means 42
The phases of the extracted waveforms WE are adjusted so that the extracted waveforms WE extracted in step 1 coincide with each other. That is, pattern matching of the extracted waveform WE is performed. For example, the extracted waveform WE extracted by the waveform extracting means 42 and the reference extracted waveform WE _ST determined by the reference extracted waveform determining means 54 or updated by the reference extracted waveform updating means 61 described later coincide with each other. And the extracted waveform W
Matching the phases of the E based extraction waveform WE _ST. Or,
At least one phase of the extracted waveforms WE is corrected so that two or more arbitrary extracted waveforms WE extracted by the waveform extracting means 42 coincide with each other. In order to match the phase of the extracted waveform WE extracted by the waveform extracting means 42 with the phase of the reference extracted waveform WE _ST , for example, the correlation coefficient between the reference extracted waveform W _ST and the extracted waveform WE is set to be maximum. The phase of the extracted waveform WE is corrected, that is, the time axis of the extracted waveform WE is shifted. Or alternatively
The phase of the extracted waveform WE is corrected so that the mean square error between the reference extracted waveform _WST and the extracted waveform WE is minimized.

The adding means 60 adds the extracted waveforms WE whose phases have been matched by the phase matching means 58. That is, the extraction waveform WE whose phases are aligned to match the reference extraction waveform W _ST in phase matching means 58 is sequentially added to the reference extraction waveform W _ST. Alternatively, two or more extracted waveforms WE whose phases are matched so that they match each other in the phase matching means 58 are added at a time. For example, if the extracted waveforms for n beats are added, the signal component generated for each beat is n times, while the randomly generated component is only √n (square root of n). When the extracted waveform WE is added by the adding means 60,
The signal strength of heart sounds and heart murmurs generated every beat is relatively increased.

The reference extraction waveform updating means 61 includes an adding means 6
The reference extraction waveform _WST is updated based on the extraction waveform WE added at 0, and a new reference extraction waveform _WST is determined. That is, a waveform obtained by dividing the amplitude intensity of the added waveform added by the adding means 60 by a number corresponding to the number of the extracted waveforms WE added by the adding means 60 is used as a new reference extracted waveform W _ST. decide.

The representative extracted waveform determining means 62 determines a representative extracted waveform WR based on the extracted waveform WE added by the adding means 60. For example, the adding means 60
Is directly determined as the representative extracted waveform WR. Or the addition means 6
A value obtained by dividing the amplitude of the extracted waveform WE after addition by 0 by the number of added extracted waveforms is determined as a representative extracted waveform WR.

The representative extracted waveform display means 64 displays the representative extracted waveform WR determined by the representative extracted waveform determination means 62 on the display 36, and the comparison display means 66 comprises a time axis and an amplitude axis. In the two-dimensional coordinate system, the extracted waveform WE extracted by the waveform extracting means 42 and the representative extracted waveform WR determined by the representative extracted waveform determining means 62 are displayed so as to be simultaneously comparable.

The time-frequency analysis means 68 outputs the representative extraction waveform WR determined by the representative extraction waveform determination means 62.
Is analyzed for time and frequency. For example, the representative extracted waveform WR is divided into a plurality of time zones, and the frequency of the representative extracted waveform WR is analyzed for each of the divided time zones. Alternatively, the representative extracted waveform WR is subjected to wavelet transform.

The wavelet transform is defined as a moving variable b for translating the wavelet function ψ (t), an example of which is shown in FIG. 5, in the time axis direction, and the magnitude of the waveform represented by the wavelet function in the time axis direction. The wavelet function ψ ((t
−b) / a) and a function f (t) representing a representative extracted waveform WR
Is defined as a function of a and b obtained by integrating the product with respect to time t. That is, it is defined as in Equation 1 below. Note that the above wavelet function ψ
In ((t−b) / a), ψ
Since the width of (t) is multiplied by a, 1 / a corresponds to the frequency, and ψ (t) moves in parallel in the time axis direction according to the moving variable b, so that b corresponds to time. .

[0034]

(Equation 1)

FIGS. 6 and 7 are diagrams for explaining the meaning of the wavelet transform equation of the above equation 1, and FIG.
Is a function g (t) having a wavelet function ψ ((t−b) / a) by appropriately selecting parameters a and b.
FIG. 7 (A) shows a state substantially matching a part of
Function h with wavelet function ψ ((t−b) / a)
A state in which a part of (t) is not approximated is shown. FIGS. 6B and 7B show the wavelet functions ウ ェ ー ブ in the cases of FIGS. 6A and 7A.
It is a figure which shows the product of ((t-b) / a) and the function g (t) or h (t), respectively. As shown in FIG. 6 (B), the wavelet function ψ ((t−b) / a) is
If they substantially coincide with a part of (t), ψ ((t−
Since the product of b) / a) and g (t) does not change sign,
The integral value increases. However, as shown in FIG. 7B, when the wavelet function ψ ((t−b) / a) does not approximate a part of the function h (t), ψ ((t−
Since the sign of the product of b) / a) and h (t) changes drastically with the change of t, the integral value decreases. Therefore, the above equation 1 is obtained by changing the parameters a and b.
When the wavelet function ψ ((t−b) / a) is similar to a part of the function f (t) representing the representative extracted waveform WR, the wavelet function 大 き い ((t−b) / a) indicates a large value, and the function f representing the representative extracted waveform WR
When it does not resemble a part of (t), it shows a small value.

The analysis result display means 70 displays the frequency analysis spectrum subjected to the time-frequency analysis by the time-frequency analysis means 68 on the display 36.

FIG. 8 is a flowchart for explaining a main part of the control operation of the electronic control unit 20. This routine is executed when the push button 28 is pressed and the activation signal SS is supplied.

First, in an initialization step (not shown), after executing a timer t for counting the pulse signal SP supplied from the clock signal source 29 and an initial process for clearing a register, a step corresponding to the waveform storage means 40 is performed. At SA <b> 1 (hereinafter, steps are omitted), the heart sound signal SH and the electrocardiogram signal SE are sequentially stored in the storage area of the RAM 34.

In the following SA2, the count content t of the timer is calculated.
Is determined whether or not 30 seconds have elapsed. Until 30 seconds have passed, the determination in SA2 is denied, and the above SA1
By repeating the following, the heart sound signal SH and the electrocardiogram signal SE are continuously stored.

If the determination at SA2 is affirmative, the control goes to SA3 corresponding to the waveform extracting means 42 to determine whether SA1
From the heart sound signal SH stored in step (1), a heart sound waveform of a predetermined section generated in each cycle is extracted as an extraction waveform WE. In SA3, the heart sound signal SH for 30 seconds is stored, that is, the heart sound waveform for 30 to 40 beats is stored, so the extracted waveform WE _n (n = 30) for 30 to 40 beats is stored.
1 to 30 to 40) are extracted. In addition, as the predetermined section, for example, the start of the section is 0.2 seconds after the occurrence of the R wave in the electrocardiogram measured as the reference lead, and the heart sound waveform represented by the heart sound signal SH for 0.4 seconds thereafter. Is extracted. 0.
0.4 seconds after 2 seconds is set as a period sufficiently including a useful II sound for diagnosis, but the section to be extracted is not limited to this section, and is within a range of one beat. Set arbitrarily.

SA4 corresponding to the following amplitude ratio calculating means 44
Now, the extracted waveform WE extracted in SA3_nEach of
The maximum amplitude D_sAnd the minimum amplitude D_nAnd the amplitude ratio (D_S
/ D _n)_nIs calculated. And the following average value calculating means
In SA5 corresponding to 46, the vibration calculated in SA4 is calculated.
Width ratio (D_S/ D_n)_nAverage value of D_AVIs calculated and the following mark
In SA6 corresponding to the quasi-deviation calculating means 48, the above SA4
The amplitude ratio (D_S/ D_n)_nStandard deviation s of
Will be issued.

Corresponding to the subsequent reference extraction waveform determining means 54
In SA7, the average value D calculated in SA5_AVFrom
Is within twice the standard deviation s calculated in SA6.
Amplitude ratio (D_S/ D_n)_nExtracted waveform WE with_nof
Of which, the amplitude ratio (D_S/ D_n) _nExtracted waveform W at which
E_nIs the first reference extraction waveform W_ST1Is determined. Sand
That is, the extracted waveform WE that satisfies Equation 2_nIn the amplitude ratio (D_S
/ D_n)_nIndicates the maximum value of the extracted waveform WE_nIs the first criterion
Extracted waveform W_ST1Is determined. It should be noted that in general,
The range includes all extracted waveforms WE_n95%.

[0043]

D _AV -2s ≦ (D _S / D _n ) _n ≦ D _AV + 2s

[0044] In SA8 corresponding to the subsequent permutation determining means 56, the order of extracted waveform WE _n used for phasing in the next SA9 is determined. That is, the order of the extracted waveform WE _n used in the next SA9 to phasing is, among the extracted waveform WE _n ranges satisfying the above Equation 2 (i.e. extracted waveform WE _n of the first range), the amplitude ratio (D _S /
D _n ) From the larger to the smaller of _n , and the above formula 1
To the amplitude ratio (D _S / D _n) towards amplitude ratio from large _n (D _S / D _n) the minimum value of _n in the filled not range of extraction waveform WE _n (i.e. extracted waveform WE _n of the second range) It is determined.

Subsequently, the number of times less than the number n of the extracted waveforms WE _n extracted in SA3 (ie, n−
SA9 to SA12 are repeated only once. First, in SA9 corresponding to the phase matching unit 58, SA7
In determined, or for a single extraction waveform WE _n the order is determined by the n-th reference extracted waveform W _STn and SA8 updated in later-described SA11, so the correlation coefficient becomes maximum, or its The phase of the extracted waveform WE _n is corrected so that the square of the difference between the amplitude intensities of the two waveforms is minimized. In SA9 corresponding to the subsequent adding means 60,
A waveform obtained by multiplying the amplitude intensity of the n-th reference extraction waveform W _STn by the number n of the extraction waveforms WE _used for calculating the reference extraction waveform W _STn is combined with the extraction waveform WE whose phase has been adjusted in SA8. _n is added.

At SA11 corresponding to the subsequent reference extraction waveform updating means 61, the waveform obtained by adding at SA10 is
Determining a waveform obtained by dividing the number (n + 1) corresponding to the number of the extracted waveform WE _n which are added by the SA10 as a new reference extraction waveform W _ST. That is, the reference extraction waveform _WST is updated. Then followed the SA12, extracted waveform WE _n extracted in the SA3 whether the summed total beats is determined. While the determination in SA12 is denied, SA9 and subsequent steps are repeatedly executed.

In the repetition of SA9 to SA12, for example, the first execution is performed in SA9. First, in SA9, the first reference extracted waveform W _ST1 determined in SA7 is obtained.
, The extracted waveform W to be added firstly determined in SA8
Phase is combined of E _n, in the subsequent SA10, the two extraction waveform is added. Then, in SA11,
The waveform obtained by dividing the amplitude intensity of the waveform obtained by adding in SA10 by 2 is a new reference extraction waveform W
Determined as _ST2 . And the second SA9 to S
In A11, firstly to the reference extraction waveform W _ST2 determined in previous SA11 in SA9, aligned phase extraction waveform WE _n to be added to the second determined in SA8, the subsequent SA10, the reference extraction waveform W extracting waveform WE _n whose phases are aligned with the reference extraction waveform W _ST2 in the SA9 is added to the amplitude intensity of _ST2 double-obtained waveform. Then, in SA11, a waveform obtained by dividing the amplitude intensity of the waveform obtained by adding in SA10 by ₃ is determined as a new reference extracted waveform W _ST3 .

As described above, SA9 to SA11 are n
−1 times, the extracted waveform WE _n extracted in SA3
Is added to all beats, the determination at SA12 is affirmed,
Finally, the reference extraction waveform W _STn determined in SA11
Is determined as a representative extracted waveform WR. Therefore, the SA 12 corresponds to the representative extracted waveform determining means 62.

At SA13 corresponding to the representative extracted waveform display means 64, the representative extracted waveform WR determined at SA12 is displayed on the display 36 and used for diagnosis of a heart disease. FIG. 9 is a diagram showing an example thereof. Since external noise is preferably removed as compared with FIG. 4 described above, it is possible to accurately diagnose a heart disease. Furthermore, in SA14 corresponding to the comparison display means 66 followed, as shown in FIG. 10, a two-dimensional coordinate composed of the time axis 72 and amplitude axis 74. Display 36, is extracted waveform WE _n extracted with SA3 is displayed in solid lines, by a representative extraction waveform WR determined in SA12 is displayed by a broken line, the extracted waveform WE _n and representative extraction waveform WR are simultaneously displayed. FIG.
0, SA is an example displayed in SA14.
Although one of the extracted waveform WE _n extracted with 3 are displayed at the same time as representative extracted waveform WR, in SA14, the two-dimensional coordinate composed of the time axis 72 and amplitude shaft 74. Display device 36, extracted with SA3 all extracted waveform WE _n which is displayed at the same time as representative extracted waveform WR is, the outer noise size as the extracted waveform WE _n from the difference between the representative extraction waveform WR, and to recognize the variation of the outer noise Is available.

S corresponding to the subsequent time-frequency analysis means 68
In A15, the representative extracted waveform WR determined in SA12 is subjected to time-frequency analysis. That is, Equation 1
, A function f (t) representing the representative extracted waveform WR is subjected to wavelet transform to obtain a moving variable (ie, time) b
And the reciprocal of the expansion variable (ie, frequency) a
(B, 1 / a).

SA corresponding to the subsequent analysis result display means 70
At 16, as shown in FIG. 11, the time axis b (that is, the moving variable axis) of the display 36 and the frequency axis (1 / a)
The magnitude of the function W (b, 1 / a) obtained by performing the wavelet transform in SA15 is displayed as a contour map on a two-dimensional plane with the axis (ie, the axis representing the reciprocal of the expansion variable). In FIG. 11, it can be seen that the signal of the high frequency component is detected at the beginning of the time axis b, and the high frequency component is weakened over time, and the signal of the low frequency component is detected instead. By comparing the frequency range in which the signal is generated, the time at which the signal is generated, or the entire display pattern of the graph with a standard pattern of a normal heart sound obtained in advance, the presence of a heart murmur and the location of the heart murmur are specified. And other diseases can be diagnosed.

As described above, according to the present embodiment, the waveform extraction
In the output means 42 (SA3), one heart rate signal SH
The heart sound waveform of the predetermined section including the II sound generated for each beat is extracted wave
Shape WE_nAnd the phase matching means 58 (SA
9) In the extracted waveform WE _nAre in phase with each other
And in the adding means 60 (SA9)
Extracted waveform WE whose phase has been matched_nIs added
With this, the normal heart sounds and murmurs that occur every beat are
Superimposed and the external noise is averaged. And teens
The tabular extraction waveform determining means 62 (SA12) uses the adding means 6
0 (SA9) based on the waveform obtained by adding
Since the representative extraction waveform WR is determined, the representative extraction waveform WR is determined.
Accurate diagnosis of heart disease using waveform WR
Wear.

According to the present embodiment, the representative extracted waveform WR is displayed by the representative extracted waveform display means 64 (SA13).
Is displayed on the display 36, so that a heart disease can be diagnosed from the displayed representative extracted waveform WR.

According to the present embodiment, the comparison display means 6
In 6 (SA14), determined in a two-dimensional coordinate system consisting of the time axis 72 and amplitude axis 74., and extracted waveform WE _n extracted by the waveform extracting means 42 (SA3) in a typical extraction waveform determining means 62 (SA12) since been a typical extraction waveform WR is comparably displayed simultaneously, it is possible to recognize the outer noise from a difference between the representative extracted waveform WR and the extracted waveform WE _n.

Further, according to the present embodiment, the amplitude ratio calculating means
At 44 (SA4), the waveform extracting means 42 (SA3)
Extracted extracted waveform WE_nAmplitude ratio (D_S/ D_n)_nBut
Each is calculated, and the average value is calculated by the average value calculating means 46 (SA5).
Average value D of the amplitude ratio of_AVIs calculated, and the standard deviation calculating means 4
8 (SA6), the standard deviation s of the amplitude ratio is calculated.
In the quasi-extracted waveform determining means 54 (SA7), the amplitude ratio (D_S
/ D_n)_nIs the average value D _AVIs twice the standard deviation s
Extracted waveform WE in the range of_nOf the amplitude ratio (D_S/
D_n)_nExtracted waveform WE that maximizes_nIs the reference extraction waveform W
_ST1Is determined as That is, such as spike noise
Amplitude ratio (D_S/ D_n)_nWE is abnormally large_n
Is the extracted waveform WE excluded_nAmong the most external noise
Extracted waveform WE with little_nIs the reference extraction waveform W_ST1Decided as
Is determined. Then, the phase matching means 58 (SA9)
Is the reference extraction waveform W_ST1Extracted waveform WE based on_n
Are extracted waveforms WE so that_nAre in phase
The extracted waveform WE with high accuracy_nPhase with each other
Can be matched. Therefore, the adding means 60 (SA1
0), the extracted waveform WE_nIs added and the representative
The addition is performed by the outgoing waveform determination means 62 (SA12).
Representative extracted waveform WR determined based on the obtained waveform
Means that heart sounds and murmurs that occur every beat
On the other hand, there are advantages to be emphasized.

Further, according to the present embodiment, the reference extraction waveform updating means 61 (S9) is provided by the phase matching means 58 (SA9).
A11) and updated reference extracted waveform W _STn phase and the phase of the previous extraction waveform WE _n to be added in the adding means 60 (SA10) is matched to each other in, in the phasing addition means 60 (SA10) Means 58 (SA
9) cross the reference extraction waveform W _STn phase was matched and extracted waveform WE _n is added to the, in the reference extraction waveform updating unit 61 (SA11), the adding means 60 (SA
The reference extracted waveform W _STn is updated based on the extracted waveform WE _n added in 10). Therefore, for each extracted waveform WE _n, the reference extraction waveform W _STn which is sequentially updated with the phase of the extracted waveform WE _n since the phase is matched to one another, one being done in the Examples which will be described next Since the phase matching accuracy is improved as compared with the case _where the phases of the extracted waveforms WEn are matched based on the reference extracted waveform _WST ,
There is an advantage that a representative extracted waveform WR in which heart sounds and murmurs are further emphasized with respect to external noises can be obtained.

Further, according to the present embodiment, the time-frequency analysis means 68 (SA15) performs the wavelet transform on the representative extracted waveform WR to obtain the time b and the frequency (1 / a).
As shown in FIG. 11, since the normal heart sound and the heart murmur included in the representative extracted waveform WR can be separated and the time when the heart murmur occurs can be analyzed, as shown in FIG. A representative extracted waveform WR such that the user can easily know the location of the heart murmur and can specify
There is an advantage that diagnosis based on the information can be easily performed.

Next, another embodiment of the present invention will be described in detail with reference to the drawings. In the following embodiments, portions common to the above-described embodiments will be denoted by the same reference numerals and description thereof will be omitted.

FIG. 12 is a functional block diagram illustrating a main part of the control operation of the electronic control unit 20 of the phonograph to which another embodiment of the present invention is applied. Note that, in the heart sound meter of this embodiment, the mechanism and circuit configuration of the device are common to those of the above-described embodiment of FIG. 1, and the control operation by the electronic control device 20 is different in the following points.

That is, in the electronic control unit 20 of this embodiment, the first reference extraction waveform W
_ST1 determines, combined all other extraction waveform WE _n phases to coincide with the reference extraction waveform W _ST1, after the combined phases of all of the extracted waveform WE _n, all of the extracted waveform W
By adding E _n at a time, to determine the representative extraction waveform WR. (Therefore, permutation determining means 56 and the reference extraction waveform updating unit 61 is not provided.) The comparison was comparably displays the extracted waveform WE _n a representative extraction waveform WR, to recognize the external noise Instead of the display means 66,
Calculating the outer noise from the extracted waveform WE _n a representative extraction waveform WR, displays and analyzes the calculated external noise. Hereinafter, the difference will be mainly described.

The external noise calculation means 82 is
The difference between the extracted waveform WE extracted by the waveform extracting unit 42 and the representative extracted waveform WR determined by the representative extracted waveform determining unit 62 in the heart sound waveform represented by the heart sound signal SH stored in the RAM 34 It is calculated as the external noise waveform WN.

The external noise display means 84 is provided by the external noise calculation means 8.
The external noise waveform WN calculated in step 2 is displayed on the display 36, and the external noise analyzing means 86 performs frequency analysis on the external noise waveform WN calculated in the external noise calculating means 82, and displays the result on the display 36. I do.

FIG. 13 is a flowchart for explaining the main part of the control operation of the electronic control unit 20 of the present embodiment. In the figure, at SB1 to SB7, the same processing as SA1 to SA7 of the above-described embodiment is performed, so that the heart sound signal SH and the electrocardiogram signal SE for 30 seconds are stored, including the II sound generated at each beat. Extracted waveform WE _n (n
= 1 to 30 to 40) are extracted, and the extracted waveform WE _n is extracted.
Among them, within the range where the deviation from the average value D _AV is twice the standard deviation s, the extracted waveform WE _n having the largest amplitude ratio (D _S / D _n ) _n is determined as the reference extracted waveform W _ST1. .

SB8 corresponding to the following phase matching means 58
So for all of the waveform of the extracted waveform WE _n extracted in SB3, the reference extraction waveform W determined in SB7
The phase is corrected so as to coincide with _ST1, and in SB9 corresponding to the subsequent adding means 60, all the extracted waveforms WE _n whose phases are corrected are added. When all the extracted waveforms WE _n are added in SB9, the normal heart sounds and heart murmurs generated for each beat are superimposed and the external noise is averaged, so that the relatively normal heart sounds and murmurs are emphasized. Is done.

[0065] followed by the SB10 corresponding to the representative extraction waveform determining means 62, by the amplitude intensity of the extracted waveform WE _n after addition calculated in SB9 is divided by the added number of beats n, the amplitude of one heartbeat is the intensity, ie amplitude are averaged, extracted waveform WE _n, which is an amplitude intensity of the one heartbeat is determined typically extracted waveform WR.

At SB11 corresponding to the representative extracted waveform display means 64, the representative extracted waveform WR determined at SB10 is displayed on the display 36, for example, as shown in FIG. corresponding to the SB12, the calculated external noise waveform WN _n for each one heartbeat from the difference between the representative extraction waveform WR determined in the extracted extracted waveform WE _n and SB10 in SB3, followed external noise display means 84 in the corresponding SB13, it is displayed outside noise waveform WN _n calculated in SB12. Figure 14 is a diagram showing an example of the external noise waveform WN _n displayed in SB13, typical extraction waveform W shown in FIG. 6 from the extraction waveform WE shown in FIG. 4
External noise waveform WN calculated by subtracting R
_n is shown. FIG. 14 shows only one external noise waveform WN _n, but in SB 13, SB 1
All external noise waveform WN _n calculated in 2 is displayed on the display unit 36, from their outer noise waveform WN _n, the size of the outer noise, and it is possible to recognize changes of the outer noise.

SB1 corresponding to the following external noise analysis means 86
At 4, the external noise waveform WN _n calculated at SB12 is subjected to Fourier analysis, and the Fourier analysis spectrum is displayed on the display 3.
6 is displayed. The external noise waveform WN _n has a heartbeat period of 4
Since it includes a plurality of noises that occur accidentally from outside the body, such as the respiratory noise having a period of up to five times, footsteps, door opening and closing sounds, and the like, these noises are frequency-separated and displayed by Fourier analysis. Is displayed on the display 36.

S corresponding to the subsequent time-frequency analysis means 68
At B15, the representative extracted waveform WR determined at SB12 is subjected to time-frequency analysis. That is, the representative extracted waveform WR is divided into a plurality of preset time zones, and frequency analysis is performed for each of the divided time zones. For example,
The time zone of the representative extracted waveform WR is divided into four equal parts, and the first section T ₁ , the second section T ₂ , the third section T ₃ , and the fourth section T 4, respectively.
And interval T _4, it is frequency analysis on the each of the four time zones.

At SB15, the representative extracted waveform W
When R is subjected to frequency analysis, the heart sound (II sound in this embodiment) and the heart murmur have different frequency components, so that the heart sound and the heart murmur are separated. In addition, since the representative extracted waveform WR is subjected to frequency analysis for each of the plurality of time zones, there are time zones in which the heart murmur is included and time zones in which the heart murmur is not included, depending on the time zone. For example, the representative extracted waveform WR in FIG.
As shown in FIG. 15, if a waveform representing the synthesized sound of the II sound and murmur, a typical extraction waveform WR of the third section T ₃ and a fourth zone T ₄ are present only in the spectrum obtained by frequency analysis The presence of the signal makes it possible to know that there is a heart murmur, and furthermore, the type of the heart murmur is discriminated from the generation time of the heart murmur, and the diseased part can be specified.

Subsequent SB corresponding to the analysis result display means 70
In FIG. 16, as shown in FIG. 16, the frequency analysis spectrum obtained as a result of the time frequency analysis in the SB 15 is displayed on the frequency axis 76 and the amplitude axis 7 of the display 36.
8, and a signal present only in the third section T ₃ and the fourth section T ₄ are displayed on the three-dimensional coordinate system including the time zone axis 80, so that the heart murmur and its generation time can be recognized. .

[0071] As described above, according to this embodiment, the waveform extracting means 42 (SB3), a heart sound signal SH, as heart sound waveform in the predetermined section is extracted waveform WE _n containing II sound generated every one heartbeat The extracted phase matching means 58 (SB
Keyed so that the phase of the extracted waveform WE _n matches the phase of the reference extraction waveform determining means 54 (SB7) reference extracted waveform W _ST1 determined in the 8), adding means 60 (SB
In 9), the extracted waveforms WE _n whose phases have been matched are all added for n beats, whereby normal heart sounds and heart murmurs generated for each beat are superimposed, and the external noise is averaged. Then, the representative extracted waveform determining means 62 (S
In B10), the representative extracted waveform WR is determined by dividing the waveform calculated by the adding means 60 (SB9) by the added number of beats n, so that the diagnosis of the heart disease is performed using the representative extracted waveform WR. Can be done accurately.

Further, according to the present embodiment, the external noise display means 84 (SB13) causes the external noise calculation means 82 (SB1
2) are calculated in, since the outer noise waveform WN _n indicating the size of the outer noise are displayed on the display unit 36, it can easily recognize the size of the outer noise, and the outer noise waveform W
N _n can be used for diagnosis.

[0073] Further, according to this embodiment, the outer noise analyzing means 86 (SB 14), an outer noise waveform WN _n is Fourier analysis, noise from respiratory noise and in vitro contained outside noise waveform WN _n Etc. are separated. Thus, using an analysis spectrum of the outer noise waveform WN _n, which is the Fourier analysis, it is possible to diagnose the disease of the respiratory system, and because it is easy to clarify the cause of the noise from outside the living body, that It is easy to remove the cause of the noise.

Further, according to the present embodiment, the time-frequency analysis means 68 (SB15) converts the representative extracted waveform WR to four.
Is divided into two time zones, and the frequency analysis is performed for each of the four time zones, so that the normal heart sound and the heart murmur included in the representative extracted waveform WR are separated, and in which time zone the heart murmur occurs Since this is analyzed, there is an advantage that diagnosis based on the representative extracted waveform WR can be easily performed. For example, it is possible to easily know the presence of the heart murmur and to specify the location where the heart murmur occurs.

While the embodiment of the present invention has been described with reference to the drawings, the present invention can be applied to other embodiments.

For example, in the above-described embodiment, in the reference extraction waveform determining means 54 (SA7, SB7), the amplitude ratio (D _s / D _n ) is within a range in which the deviation from the average value D _AV is twice the standard deviation s. in the extracted waveform WE _n located within, the amplitude ratio (D
_s / D _n) but has been determined extracted waveform WE _n having a maximum relative extracted waveform W _ST, not calculated amplitude ratio (D _s / D _n) is a waveform extracting unit 42 in (SA3, SB3) it one any of the extracted extracted waveform WE _n may be determined based on the extraction waveform W _ST, the reference extraction waveform W
The phase may be corrected so that the extracted waveforms WE _n coincide with each other without determining the _ST . Reference so even if the case any extraction waveform WE _n is determined based on the extraction waveform W _ST or reference extracted waveform W _ST is not determined, the reference extraction waveform W _ST in phase alignment means 58 (SA9, SB8) as it is modified other extraction waveform WE _n phases, or the phase of the extracted waveform WE _n are modified to cross the phases coincide, normal heart sounds and murmurs by addition is superimposed, the outer noise A tentative effect on averaging is obtained.

In the first embodiment, the extracted waveform WE _{n is} output at SA 14 corresponding to the comparison display means 66.
Are displayed as a broken line and the representative extracted waveform WR is displayed as a solid line, so that the two can be compared. However, even if the comparison can be made by other display methods such as displaying using different colors. Good.

The electrocardiograph of the above-described embodiment has an electrode 22, an electrocardiographic signal amplifier 24, and an A / D signal so that an electrocardiogram can be measured as a reference lead in addition to the measurement of the electrocardiogram.
Although a D converter 26 was provided, the electrocardiogram need not necessarily be measured. When the electrocardiogram is not measured as the reference lead, the waveform extracting means 42 outputs the heart sound signal S
One or two predetermined portions occurring every one beat of H are directly detected, and a predetermined section or a waveform between the two predetermined portions is extracted from the predetermined portion.

In the first embodiment, the size of the function W (b, 1 / a) obtained by the wavelet transform in the analysis result display means 70 (SA16) is:
Although a contour map is displayed on a two-dimensional plane of the time axis b and the frequency axis 1 / a on the display 36, a tertiary map formed by the time axis b, the frequency axis 1 / a, and the axis representing the conversion value The original coordinates may be displayed as a three-dimensional graph.

In addition, the present invention can be variously modified without departing from the gist thereof.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a heart sound meter as one embodiment of the present invention.

FIG. 2 is a functional block diagram illustrating a main part of a control function of the electronic control device of the embodiment of FIG. 1;

FIG. 3 is a diagram showing an example of a heart sound diagram and an electrocardiogram represented by a heart sound signal and an electrocardiogram signal.

FIG. 4 is a diagram illustrating an example of an extracted waveform extracted by a waveform extracting unit.

FIG. 5 is a diagram illustrating a wavelet function ψ (t).

FIG. 6 is a diagram illustrating a state where the wavelet function ψ ((t−b) / a) substantially matches a part of a certain function g (t), and a product of the two functions at that time.

FIG. 7 is a diagram illustrating a state in which the wavelet function ψ ((t−b) / a) does not approximate a part of a certain function h (t), and a product of the two functions at that time.

FIG. 8 is a flowchart illustrating a main part of a control operation of the electronic control device according to the embodiment of FIG. 1;

FIG. 9 is a diagram showing an example of a representative extracted waveform displayed by a representative extracted waveform display means.

FIG. 10 is a diagram showing an example in which an extracted waveform and a representative extracted waveform are displayed on a display by a comparison display means.

FIG. 11 is a contour diagram displayed on a display unit in the analysis result display means of the embodiment of FIG. 1;

FIG. 12 is a functional block diagram illustrating a main part of a control function of an electronic control device according to another embodiment of the present invention.

FIG. 13 is a flowchart illustrating a main part of a control operation of the electronic control device of the embodiment in FIG. 10;

FIG. 14 is a diagram showing an example of an external noise waveform displayed on a display in the external noise display means.

FIG. 15 is a diagram for explaining that a time zone in which a heart murmur occurs can be specified by the time-frequency analysis means.

FIG. 16 is a diagram showing an example of a frequency analysis spectrum displayed on the analysis result display means.

[Description of sign]

 10: Heart sound meter 12: Microphone (heart sound sensor) 42: Waveform extraction means 44: Amplitude ratio calculation means 46: Average value calculation means 48: Standard deviation calculation means 54: Reference extraction waveform determination means 58: Phase matching means 60: Addition means 61: reference extraction waveform updating means 62: representative extraction waveform determination means 64: representative extraction waveform display means 66: comparison display means 68: time frequency analysis means 82: external noise calculation means 84: external noise display means 86: external noise Analysis means

Claims (2)

[Claims] 1. A heart sound meter for detecting a heart sound generated from a heart and outputting a heart sound signal representing the heart sound, wherein the heart sound sensor measures a heart sound waveform represented by the heart sound signal. Waveform extracting means for extracting a heart sound waveform of a predetermined section generated every time as an extracted waveform; phase matching means for adjusting the phases of the extracted waveforms so that the extracted waveforms extracted by the waveform extracting means coincide with each other; Adding means for adding the extracted waveforms whose phases have been matched by the phase matching means; and representative extracted waveform determining means for determining a representative extracted waveform based on the extracted waveform calculated by the adding means. And a heart sound meter. 2. The electrocardiograph according to claim 1, further comprising an external noise calculating means for calculating an external noise waveform from a difference between the extracted waveform extracted by the waveform extracting means and the representative extracted waveform.