JP6467044B2 - Shunt sound analysis device, shunt sound analysis method, computer program, and recording medium - Google Patents

Description

  The present invention relates to a technical field of a shunt sound analysis device, a shunt sound analysis method, a computer program, and a recording medium that analyze a shunt sound acquired from a measurement subject.

  As this type of device, there is known a device that analyzes a shunt sound obtained from a measurement subject and assists a doctor's diagnosis regarding shunt stenosis or the like. For example, Patent Document 1 describes a technique for presenting a degree of stenosis corresponding to a peak position of an envelope component from which a fine component has been removed by utilizing the fact that the position of a peak of frequency characteristics corresponds to the degree of stenosis.

JP 2010-29434 A

  In the technique described in Patent Document 1 described above, the degree of stenosis of the blood vessel is determined using a high-frequency component (that is, a component having a relatively high frequency) of the shunt sound. However, a characteristic timbre (for example, wind noise, etc.) that does not necessarily depend on the level of the frequency is also generated at the time of constriction. In particular, it is known that such a timbre component is also generated by factors other than stenosis, such as a temporary thrombus and the influence of a venous valve. Therefore, the technique of Patent Document 1 that does not take into account characteristic timbre components causes a technical problem that the degree of stenosis cannot be accurately evaluated.

  The characteristic timbre component depends on the degree of stenosis. A harmonic peak appears superimposed on the envelope component of the frequency characteristic and fluctuates with time. Such a characteristic cannot be expressed by the method using the frequency characteristic peak of Patent Document 1, so even if the above-described timbre component is considered, it is easily used in the conventional method. I can't.

  Examples of problems to be solved by the present invention include the above. The present invention provides a shunt sound analysis apparatus, a shunt sound analysis method, a computer program, and a recording medium that can suitably support a diagnosis of stenosis at a shunt formation site by analyzing a shunt sound acquired from a measurement subject. This is the issue.

  A shunt sound analyzing apparatus for solving the above-described problems is caused by stenosis of a blood vessel of the measurement subject from the acquisition means for acquiring shunt sound information related to the shunt formation site of the measurement subject and a shunt sound information of the measurement subject. Extraction means for extracting a characteristic component indicating the characteristic of the shunt sound, and output means for outputting evaluation information related to the evaluation of the shunt sound based on the characteristic component.

  A shunt sound analysis method for solving the above-described problem is caused by acquiring a shunt sound information relating to a shunt sound of a shunt formation site of the measurement subject and by narrowing of the blood vessel of the measurement subject from the shunt sound information. An extraction step of extracting a feature component indicating the feature of the shunt sound; and an output step of outputting evaluation information related to the evaluation of the shunt sound based on the feature component.

  A computer program for solving the above-described problems includes an acquisition step of acquiring shunt sound information related to a shunt formation site of a measurement subject, and the shunt generated by stenosis of the measurement subject's blood vessel from the shunt sound information. The computer executes an extraction step of extracting a feature component indicating a feature of the sound and an output step of outputting evaluation information related to the evaluation of the shunt sound based on the feature component.

  The above-described computer program is recorded on a recording medium for solving the above problems.

It is a block diagram which shows the whole structure of the shunt sound analyzer which concerns on an Example. It is a wave form diagram which shows an example of a shunt sound waveform. It is a spectrogram which shows an example of a time frequency waveform. It is a wave form diagram which shows an example of the frequency waveform in the time n. It is a wave form diagram which shows an example of the characteristic component characteristic waveform in the time n. It is a spectrogram which shows an example of a time characteristic component characteristic waveform. It is a block diagram which shows the specific structure of a parameter calculating part. It is a time chart which shows an example of a harmonic component value, a frequency gravity center value, and a timbre tendency value. It is a comparison list figure which shows the calculation result in a parameter calculation part for every some shunt sound. It is a top view which shows the example of a display in a display part (the 1). It is a top view (the 2) which shows the example of a display in a display part. It is a top view (the 3) which shows the example of a display in a display part. It is a top view (the 4) which shows the example of a display in a display part.

<1>
The shunt sound analysis apparatus according to the present embodiment includes an acquisition unit that acquires shunt sound information related to a shunt formation site of the measurement subject, and the shunt generated by stenosis of the measurement subject's blood vessel from the shunt sound information. Extraction means for extracting a feature component indicating the feature of the sound, and output means for outputting evaluation information related to the evaluation of the shunt sound based on the feature component.

  During the operation of the first shunt sound analyzer according to the present embodiment, shunt sound information related to the shunt sound is first acquired from the periphery of the measurement subject's shunt formation region by the acquisition means. Here, the “shunt sound” is a blood flow sound acquired in the vicinity of a shunt formation site for taking blood out of the body, and is a sound synchronized with the pulse of the measurement subject. The acquisition of the shunt sound may be performed using various sensors, and the acquisition method is not particularly limited. The “shunt sound information” is information including various parameters related to the shunt sound, and includes, for example, temporal changes such as volume and frequency.

  When the shunt sound information is acquired, the characteristic component is extracted from the shunt sound information by the extraction unit. The characteristic component is a component indicating the characteristics of the shunt sound caused by the narrowing of the blood vessel of the measurement subject. For example, from the time-frequency waveform indicating the shunt sound information, the frequency range of the envelope component of the frequency characteristics and the frequency range of the pitch component specific to the shunt sound This is a product obtained by removing fine components exceeding.

  When the feature component is extracted, evaluation information related to the evaluation of the shunt sound is output by the output means. Specifically, the output unit generates evaluation information based on the extracted feature component, and outputs the generated evaluation information to an external monitor or the like. Here, the “evaluation information” may be information indicating a result of evaluating the degree of stenosis at the shunt formation site or information for evaluating the degree of stenosis at the shunt formation site. More specifically, a numerical value or the like that directly indicates the degree of stenosis may be output as evaluation information, or one or more parameters for a doctor or the like to determine the degree of stenosis may be output as evaluation information. .

  As described above, the evaluation information according to the present embodiment is based on the characteristic component extracted from the shunt sound information. In particular, when a stenosis occurs in a blood vessel at a shunt formation site, the high frequency component of the shunt sound tends to be strong. It has been found that a timbre component also occurs. For example, when a stenosis occurs, an acoustic component expressed by an artificial sound such as Hugh or view, or a component expressed by an artificial sound such as Shanshan or Shash.

  According to the shunt sound analysis apparatus according to the present embodiment, the evaluation information considering the characteristic timbre component described above is output. Therefore, for example, more accurate evaluation can be realized as compared with the case where the stenosis is evaluated only by the frequency of the shunt sound. Therefore, it is possible to favorably support the diagnosis of stenosis at the shunt formation site.

<2>
In one aspect of the shunt sound analysis apparatus according to the present embodiment, the extraction unit performs a cepstrum conversion process on the shunt sound information, and removes a cefency of the first order or less and a cefency of the second order or more. The feature component is extracted by Fourier transform.

  According to this aspect, when shunt sound information is acquired, first, cepstrum conversion processing is performed. The cepstrum transformation process is realized, for example, by performing inverse Fourier transformation on a time frequency waveform indicating shunt sound information.

  Subsequently, from the cepstrum-transformed shunt sound information, the quefrency of the first order or less and the quefrency of the second order or more are removed. Here, the “second order” is an order set for removing fine components exceeding the frequency of the pitch component peculiar to the shunt sound, and an optimum value is set by a prior simulation or the like. Further, the “first order” is an order set to remove the envelope component of the frequency characteristic of the shunt sound (for example, the characteristic indicated by the waveform indicating the magnitude of the shunt sound for each frequency), Similar to the first order, the optimum value is set by a prior simulation or the like. Note that the optimum values of the “first order” and the “second order” may be values according to the sampling frequency and the frame length.

  The shunt sound information from which the quefrency according to the first order and the second order has been removed is converted to the original frequency waveform by Fourier transform. The waveform obtained in this way is obtained by extracting a characteristic timbre component (that is, a characteristic component) generated when stenosis occurs.

  As described above, the characteristic components of the shunt sound information can be suitably extracted by using the cepstrum transformation and the predetermined order of quefrency removal.

<3>
In another aspect of the shunt sound analysis apparatus according to the present embodiment, the output unit calculates a harmonic component value indicating a cross-correlation between adjacent frames on the time axis in the feature component, and outputs the calculated harmonic component value as the evaluation information. To do.

  According to this aspect, when the feature component is extracted from the shunt sound information, the harmonics indicating the cross-correlation between frames adjacent to the feature component on the time axis (for example, the frame at time n−1 and the frame at time n). Sex component values are calculated. The harmonic component value calculated in this way is a parameter indicating a component amount that is sustained in the time direction in the shunt sound.

  According to the study by the present inventor, it has been found that the characteristic timbre generated when stenosis occurs has a continuous component in the time direction. Therefore, if the harmonic component value is calculated, the influence of the instantaneous (that is, non-sustaining) characteristic component is reduced, and a value corresponding to the characteristic tone color component magnitude (intensity) can be obtained. . Therefore, if the harmonic component value is output as evaluation information, the occurrence of stenosis can be suitably evaluated.

<4>
In another aspect of the shunt sound analysis apparatus according to the present embodiment, the output means calculates a frequency centroid value of the feature component and outputs it as the evaluation information.

  According to this aspect, when the feature component is extracted from the shunt sound information, the frequency centroid value of the feature component is calculated. The frequency centroid value is a value corresponding to the height of the characteristic timbre component, and is calculated as a value that varies depending on, for example, the degree of stenosis and other factors. Therefore, if the frequency centroid value is output as evaluation information, the occurrence of stenosis can be suitably evaluated.

<5>
In another aspect of the shunt sound analysis apparatus according to the present embodiment, the output means calculates a timbre tendency value obtained by subtracting the average value of the feature component from the peak value of the feature component, and outputs it as the evaluation information.

  According to this aspect, when a feature component is extracted from the shunt sound information, first, a peak value (maximum peak value for a predetermined time) and an average value (average value for a predetermined time) of the feature component are calculated. Then, the average value is subtracted from the peak value of the characteristic component, and calculated as a timbre tendency value.

  The timbre tendency value is a value indicating which tendency is strong, for example, an acoustic component such as hugh or a noisy component such as a shash. Specifically, when the timbre tendency value is high, it can be determined that the tendency of the acoustic component is strong, and when the timbre tendency value is low, it can be determined that the tendency of the noise component is strong. Therefore, if the timbre tendency value is output as evaluation information, the occurrence of stenosis can be suitably evaluated.

<6>
The shunt sound analysis method according to the present embodiment includes an acquisition step of acquiring shunt sound information relating to a shunt formation site of a measurement subject's shunt formation, and the shunt caused by stenosis of the measurement subject's blood vessel from the shunt sound information. An extraction step of extracting a feature component indicating a feature of the sound; and an output step of outputting evaluation information related to the evaluation of the shunt sound based on the feature component.

  According to the shunt sound analysis method according to the present embodiment, the evaluation information based on the characteristic component generated by the stenosis is output, as in the shunt sound analysis apparatus according to the present embodiment described above. Therefore, it is possible to favorably support the diagnosis of stenosis at the shunt formation site.

<7>
The computer program according to the present embodiment obtains shunt sound information related to the shunt sound of the measurement subject's shunt formation site, and the shunt sound generated by stenosis of the subject's blood vessel from the shunt sound information. The computer executes an extraction step of extracting a feature component indicating a feature and an output step of outputting evaluation information related to the evaluation of the shunt sound based on the feature component.

  According to the computer program which concerns on this embodiment, each process of the shunt sound analysis method which concerns on this embodiment mentioned above can be performed by a computer. Therefore, it is possible to favorably support the diagnosis of stenosis at the shunt formation site.

<8>
The recording medium according to the present embodiment records the above-described computer program.

  According to the recording medium according to the present embodiment, it is possible to output evaluation information based on a characteristic component generated by stenosis by executing a recorded computer program. Therefore, it is possible to favorably support the diagnosis of stenosis at the shunt formation site.

  The operation and other gains of the shunt sound analysis apparatus, shunt sound analysis method, computer program, and recording medium according to the present embodiment will be described in more detail in the following examples.

  Hereinafter, embodiments of the shunt sound analyzer will be described in detail with reference to the drawings.

<Device configuration>
First, the overall configuration of the shunt sound analyzer according to the present embodiment will be described with reference to FIG. FIG. 1 is a block diagram illustrating the overall configuration of the shunt sound analysis apparatus according to the embodiment.

  In FIG. 1, the shunt sound analysis apparatus according to the present embodiment includes a shunt sound input unit 110, an audio signal analysis processing unit 120, a parameter calculation unit 130, an evaluation information calculation output unit 140, and a display unit 150. Configured.

  The shunt sound input unit 110 is a specific example of “acquisition means”. For example, the shunt sound input from a vibration sensor or the like performs analog-digital conversion at a predetermined sampling frequency Fs, and outputs a shunt sound waveform. The shunt sound waveform generated by the shunt sound input unit 110 is output to the audio signal analysis processing unit 120.

  The audio signal analysis processing unit 120 performs various types of analysis processing (for example, short-time Fourier transform, logarithmic transformation, etc.) on the shunt sound waveform input from the shunt sound input unit 110, and outputs a time-frequency analysis waveform. The analysis result by the audio signal analysis processing unit 120 is output to the parameter calculation unit 130.

  The parameter calculation unit 130 is a specific example of “extraction means” and performs various calculation processes (for example, inverse Fourier transform, quefrency cut, Fourier transform) on the time-frequency analysis waveform input from the audio signal analysis processing unit 120. Etc.) to obtain a time characteristic component characteristic waveform. The parameter calculation unit 130 calculates a harmonic component value, a frequency centroid value, and a timbre tendency value from the time characteristic component characteristic waveform. The harmonic component value, the frequency centroid value, and the timbre tendency value calculated by the parameter calculation unit 130 are output to the evaluation information calculation output unit 140, respectively.

  The evaluation information calculation output unit 140 is a specific example of the “output unit”, and the degree of stenosis of the blood vessel of the measurement subject is determined using the output value of the parameter calculation unit 130 and the peak value or time average value of the output value. Calculate and output evaluation information for evaluation. The evaluation information calculated by the evaluation information calculation output unit 140 is output to the display unit 150.

  The display unit 150 is configured as a monitor, for example, and is configured to be able to visually present the evaluation information input from the evaluation information calculation output unit 140 to a device user such as a doctor.

<Description of operation>
Next, a specific operation of the shunt sound analyzer according to the present embodiment will be described. In the following, among the parts of the shunt sound analysis apparatus according to the present embodiment, the parts unique to the present embodiment (specifically, the audio signal analysis processing unit 120, the parameter calculation 130, the evaluation information calculation output unit) 140 and the operation of the display unit 150) will be described in detail.

<Audio signal analysis processing unit>
First, the operation of the audio signal analysis processing unit 120 will be described in detail with reference to FIGS. 2 and 3. FIG. 2 is a waveform diagram showing an example of a shunt sound waveform. FIG. 3 is a spectrogram showing an example of a time-frequency waveform.

  As shown in FIG. 2, the shunt sound waveform generated by the shunt sound input unit 110 is input to the audio signal analysis processing unit 120. As can be seen from the figure, the shunt sound waveform is a waveform indicating the temporal variation of the amplitude of the voice acquired from the measurement subject's shunt sound part.

  The audio signal analysis processing unit 120 performs N-point short-time Fourier transform and logarithmic transformation on the value X (n) of the time n of the shunt sound waveform in units of length N frames, and generates a time-frequency analysis waveform Plog. [N, k] is calculated.

  Specifically, the audio signal analysis processing unit 120 performs short-time Fourier transform using the following mathematical formula (1).

また音声信号解析処理部１２０は、以下の数式（２）を用いて、対数変換を行う。

The audio signal analysis processing unit 120 performs logarithmic conversion using the following mathematical formula (2).

なお、ｗ（ｋ）は、フレームを切り出す際に用いられる長さＮの窓関数である。またｋは、サンプリング周波数ＦｓのＮ等分を単位とした周波数の位置（高さ）を表しており、ｋ＝０…Ｎ−１である。

Note that w (k) is a window function of length N used when cutting out a frame. K represents the position (height) of the frequency in units of N equal to the sampling frequency Fs, and k = 0... N-1.

  As shown in FIG. 3, the time-frequency analysis waveform Plog [n, k] calculated by the audio signal analysis processing unit 120 is a waveform indicating the temporal variation of the sound pressure corresponding to each frequency.

<Parameter calculation unit>
Next, the operation of the parameter calculation unit 130 will be described in detail with reference to FIGS. 4 to 9.

<Extraction of characteristic components>
First, feature component extraction by the parameter calculation unit 130 will be described in detail with reference to FIGS. 4 to 6. FIG. 4 is a waveform diagram showing an example of a frequency waveform at time n, and FIG. 5 is a waveform diagram showing an example of a characteristic component characteristic waveform at time n. FIG. 6 is a spectrogram showing an example of a time characteristic component characteristic waveform.

  As shown in FIG. 4, the frequency waveform at time n of the time-frequency analysis waveform Plog [n, k] is represented as a waveform indicating the sound pressure for each frequency. The parameter calculation unit 130 performs various calculation processes on each frame of such a time-frequency analysis waveform Plog [n, k], and extracts feature components.

  Specifically, the parameter calculation unit 130 performs inverse Fourier transform on the time-frequency analysis waveform Plog [n, k] using the following mathematical formula (3).

次に、パラメータ演算部１３０は、以下の数式（４）を用いて、ｌｉｆｔ＿ｌｏｗ次以下、及びｌｉｆｔ＿ｈｉｇｈ次以上のケフレンシーをカットする。

Next, the parameter calculation unit 130 uses the following formula (4) to cut the quefrencies below the lift_low order and above the lift_high order.

ｌｉｆｔ＿ｌｏｗ次以下のケフレンシーをカットすることで、各フレームの周波数特性の包絡成分を除去することが可能である。また、ｌｉｆｔ＿ｈｉｇｈ次以上のケフレンシーをカットすることで、シャント音特有のピッチ成分の周波数範囲を超えた微細成分を除去することが可能である。

It is possible to remove the envelope component of the frequency characteristic of each frame by cutting the cefency below the lift_low order. In addition, by cutting the cefency higher than the lift_high order, it is possible to remove fine components that exceed the frequency range of the pitch component peculiar to the shunt sound.

  Note that the optimum values of the orders lift_low and lift_high are values determined according to the sampling frequency Fs and the frame length N, and the optimum values are set by a prior simulation or the like. The order lift_low is a specific example of “first order”, and the order lift_high is a specific example of “second order”.

  Next, the parameter calculation unit 130 performs Fourier transform using the following formula (5), and calculates the time feature component characteristic waveform FPeaks [n, k].

図５に示すように、時間特徴成分特性波形ＦＰｅａｋｓ［ｎ，ｋ］の時刻ｎにおける波形は、周波数波形から特徴成分を抽出した波形として表される。

As shown in FIG. 5, the waveform at time n of the time feature component characteristic waveform FPeaks [n, k] is represented as a waveform obtained by extracting the feature component from the frequency waveform.

  As shown in FIG. 6, the time characteristic component characteristic waveform FPeaks [n, k] is a waveform showing temporal variation of the fine component of the sound pressure corresponding to each frequency.

<Calculation of characteristic component parameters>
Next, the calculation of the characteristic component parameter by the parameter calculation unit 130 will be described in detail with reference to FIGS. FIG. 7 is a block diagram showing a specific configuration of the parameter calculation unit. FIG. 8 is a time chart showing examples of harmonic component values, frequency centroid values, and timbre tendency values.

  In FIG. 7, the time characteristic component characteristic waveform FPeaks [n, k] is input to the harmonic component calculator 131, the frequency centroid calculator 132, and the timbre tendency calculator 133 of the parameter calculator 130, respectively.

  The harmonic component calculation unit 131 calculates a harmonic component value FPeaksValue [n], which is one of the characteristic component parameters, from the time characteristic component characteristic waveform FPeaks [n, k] using the following formula (6). calculate.

調波性成分値ＦＰｅａｋｓＶａｌｕｅ［ｎ］は、前後フレームの特徴成分であるＦＰｅａｋｓ［ｎ−１，ｋ］とＦＰｅａｋｓ［ｎ，ｋ］との相互相関値であり、時間方向に持続的な成分量を示している。

The harmonic component value FPeaksValue [n] is a cross-correlation value between FPeaks [n−1, k] and FPeaks [n, k], which are characteristic components of the preceding and succeeding frames, and has a sustained component amount in the time direction. Show.

  The frequency center-of-gravity calculation unit 132 calculates a frequency center-of-gravity value FPeaksCentroid [n], which is one of the feature component parameters, from the temporal feature component characteristic waveform FPeaks [n, k] using the following formula (7).

周波数重心値ＦＰｅａｋｓＣｅｎｔｒｏｉｄ［ｎ］は、狭窄の度合いやその他の要因によって異なる値として算出される。

The frequency centroid value FPpeaks Centroid [n] is calculated as a different value depending on the degree of stenosis and other factors.

  The timbre tendency calculation unit 133 calculates a timbre tendency value FPeaksSoundColor [n], which is one of the characteristic component parameters, from the temporal characteristic component characteristic waveform FPeaks [n, k] using the following formula (8).

音色傾向値ＦＰｅａｋｓＳｏｕｎｄＣｏｌｏｒ［ｎ］は、特徴成分ＦＰｅａｋｓ［ｎ，ｋ］の最大ピークから、ＦＰｅａｋｓ［ｎ，ｋ］の平均値を差し引くことで算出される。

The timbre tendency value FPeaksSoundColor [n] is calculated by subtracting the average value of FPeaks [n, k] from the maximum peak of the characteristic component FPeaks [n, k].

  As shown in FIG. 8, the harmonic component value FPeaksValue [n] is calculated as a value corresponding to the characteristic tone color component size. The frequency centroid value FPpeaks Centroid [n] is calculated as a value corresponding to the height of the characteristic timbre component. The timbre tendency value FPpeaksSoundColor [n] is calculated as a value indicating the timbre tendency of the characteristic timbre component.

<Evaluation using calculation results>
Next, a method for evaluating the degree of stenosis using various parameters obtained by the above-described calculation will be specifically described with reference to FIG. FIG. 9 is a comparative list showing the calculation results in the parameter calculation unit for each of a plurality of shunt sounds.

  In FIG. 9, a shunt sound A is an example of a normal shunt sound. The shunt sound A does not show a characteristic characteristic as can be seen from the time characteristic component characteristic waveform FPeaks [n, k]. Therefore, all of the harmonic component value FPeaksValue [n], the frequency centroid value FPeaksCentroid [n], and the timbre tendency value FPeaksSoundColor [n] are stable at extremely low values. In this way, when the characteristic component parameter is stable at a low value, it can be evaluated that stenosis has not occurred (or the degree of stenosis is extremely small).

  The shunt sound B is an example of a shunt sound including a distinctive sound. Compared with the shunt sound A, the shunt sound B is calculated as a higher harmonic component value FPeaksValue [n]. Further, the frequency centroid value FPeaksCentroid [n] also changes at a high value, and it can be seen that the height fluctuates in synchronization with the heartbeat. From the timbre tendency value FPpeaksSoundColor [n], it can be seen that there is a strong noise tendency. As a result, the shunt sound B includes a noisy characteristic component, and it can be evaluated that there is a possibility that stenosis has occurred.

  The shunt sound C is an example of a shunt sound including a characteristic sound called view view. The shunt sound C is calculated as a harmonic component value FPeaksValue [n] having a very high value. Further, the frequency centroid value FPeaksCentroid [n] changes at an extremely high value. It can be seen from the timbre tendency value FPpeaksSoundColor [n] that the acoustic tendency is strong. As a result, it can be evaluated that the shunt sound C includes a lot of acoustic characteristic components and is highly likely to have stenosis.

  The shunt sound D is an example of a shunt sound including a characteristic sound such as a low beat. The shunt sound D is calculated with a high harmonic component value FPpeaksValue [n]. Further, the frequency centroid value FPeaksCentroid [n] changes intermittently and violently. It can be seen from the timbre tendency value FPpeaksSoundColor [n] that the acoustic tendency is strong. As a result, the shunt sound C includes an acoustic feature component, and it can be evaluated that there is a possibility that stenosis has occurred.

  Note that the evaluation method described above is merely an example, and different evaluations may be performed from the same calculation result.

<Evaluation information calculation output unit>
Next, the operation of the evaluation information calculation output unit 140 will be described in detail.

  The evaluation information calculation output unit 140 quantifies each of the harmonic component value FPeaksValue [n], the frequency centroid value FPeaksCentroid [n], and the timbre tendency value FPeaksSoundColor [n] calculated by the parameter calculation unit 130 to narrow the stenosis. Output as evaluation information for evaluating the degree (or indicating the degree of stenosis).

  Specifically, the evaluation information calculation output unit 140 normalizes the harmonic component value FPeaksValue [n] to 0 to 100 in the range of 0 dB to 20 dB and outputs the normalized value. Further, the frequency centroid value FPeaksCentroid [n] is normalized to 0 to 100 in the range of 150 Hz to 2 kHz and output. Further, the evaluation information calculation output unit 140 calculates and outputs a peak value and a time average value for each of the harmonic component value FPeaksValue [n] and the frequency centroid value FPeaksCentroid [n].

  The evaluation information calculation output unit 140 normalizes the timbre tendency value FPpeaksSoundColor [n] to 0 to 100 in the range of 0 dB to 10 dB, and outputs the normalized value. Further, the evaluation information calculation output unit 140 calculates and outputs a peak value and a time average value for the timbre tendency value FPpeaksSoundColor [n].

<Display section>
Next, the operation of the display unit 150 will be described in detail with reference to FIGS. 10 to 13. 10 to 13 are plan views showing display examples on the display unit.

  As shown in FIGS. 10 and 11, the display area 155 of the display unit 150 includes, for example, an average score (that is, normalized time) of the feature component size (that is, harmonic component value FPeaksValue [n]). Average value) and an impression map for the average score are displayed. Note that the average score here is a parameter that can be evaluated as high possibility that stenosis has occurred when the average score is 30 or more.

  Specifically, in the example of FIG. 10, an average score “5” and an impression map in which almost no feature component appears are displayed. In this case, for example, it is unlikely that stenosis has occurred, and it can be determined that there is no problem with dialysis on the day. On the other hand, in the example of FIG. 11, an average score “70” and an impression map in which feature components appear strongly are displayed. In this case, for example, it is highly possible that stenosis has occurred, and it can be determined that a doctor needs to be contacted. Such a relatively simple display mode is effective when it is difficult to determine how to hear the shunt sound during auscultation.

  As shown in FIGS. 12 and 13, the display area 155 of the display unit 150 includes, for example, each of the harmonic component value FPeaksValue [n], the frequency centroid value FPeaksCentroid [n], and the timbre tendency value FPeaksSoundColor [n]. A peak value and an average score, and an analysis waveform may be displayed. If the evaluation information is displayed in detail in this way, a more accurate and detailed diagnosis (for example, diagnosis of whether the increase in score is caused by stenosis or the influence of other factors) can be performed.

  Specifically, in the example of FIG. 12, since the characteristic timbre fluctuates and is intermittent, it can be determined that stenosis should be suspected. Further, in the example of FIG. 13, a characteristic sound is seen, but since the sound seems to be continuous, it can be determined that other findings should be diagnosed together.

  The display unit 150 may perform display in a mode other than the display example described above. Moreover, it may be configured such that the user of the apparatus can appropriately select an appropriate display mode.

  As described above, according to the shunt sound analysis apparatus according to the present embodiment, appropriate evaluation information is output based on the acquired shunt sound information. Therefore, it is possible to favorably support the diagnosis of stenosis at the shunt formation site.

  The present invention is not limited to the above-described embodiments, and can be appropriately changed without departing from the spirit or idea of the invention that can be read from the claims and the entire specification, and shunt sound analysis accompanying such changes is possible. A device, a shunt sound analysis method, a computer program, and a recording medium are also included in the technical scope of the present invention.

DESCRIPTION OF SYMBOLS 110 Shunt sound input part 120 Audio | voice signal analysis process part 130 Parameter calculation part 131 Harmonic component calculation part 132 Frequency gravity center calculation part 133 Tone tendency value calculation part 140 Evaluation information calculation output part 150 Display part 155 Display area FPeaksValue [n] key Wave component value FPeaks Centroid [n] Frequency centroid value FPeaksSoundColor [n] Tone tendency value

Claims (7)

An acquisition means for acquiring shunt sound information related to the shunt sound of the shunt formation site of the measurement subject;
A cepstrum conversion process is performed on the shunt sound information, and the cuff frequency of the first order or lower and the cef frequency of the second order or higher are removed and Fourier-transformed, whereby the shunt sound generated by the stenosis of the blood vessel of the measurement subject. Extracting means for extracting a feature component indicating the characteristics of
An output unit that outputs evaluation information related to the evaluation of the shunt sound based on the characteristic component. 2. The shunt sound analysis according to claim 1 , wherein the output unit calculates a harmonic component value indicating a cross-correlation between adjacent frames on the time axis in the feature component, and outputs the harmonic component value as the evaluation information. apparatus. Said output means, said calculating a frequency centroid value of the characteristic components, shunt sound analysis apparatus according to claim 1 or 2, characterized in that output as the evaluation information. Said output means calculates the tone tendency value obtained by subtracting the average value of the characteristic component from the peak value of the characteristic component, in any one of claims 1 to 3, characterized in that output as the evaluation information The shunt sound analyzer described. An acquisition step of acquiring shunt sound information related to the shunt formation site of the measurement subject's shunt formation;
A cepstrum conversion process is performed on the shunt sound information, and the cuff frequency of the first order or lower and the cef frequency of the second order or higher are removed and Fourier-transformed, whereby the shunt sound generated by the stenosis of the blood vessel of the measurement subject. An extraction process for extracting feature components indicating the characteristics of
An output step of outputting evaluation information related to the evaluation of the shunt sound based on the characteristic component. An acquisition step of acquiring shunt sound information related to the shunt formation site of the measurement subject's shunt formation;
A cepstrum conversion process is performed on the shunt sound information, and the cuff frequency of the first order or lower and the cef frequency of the second order or higher are removed and Fourier-transformed, whereby the shunt sound generated by the stenosis of the blood vessel of the measurement subject. An extraction process for extracting feature components indicating the characteristics of
A computer program for causing a computer to execute an output step of outputting evaluation information related to the evaluation of the shunt sound based on the characteristic component. A recording medium in which the computer program according to claim 6 is recorded.

Images