JP7303921B2 - 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 for analyzing shunt sound obtained from a subject.

As this type of device, there is known a device that analyzes the shunt sound acquired from the subject and assists the doctor in diagnosing shunt stenosis or the like. For example, in Patent Document 1, an array-shaped sound pickup sensor is fixed to the arm of a person to be measured, and shunt sounds are acquired from a plurality of locations. A shunt sound analysis apparatus is disclosed for performing analysis using the STMEM method or the like.

JP 2014-8263 A

However, in the shunt sound analysis apparatus as described in the above-mentioned Patent Document 1, not only is it necessary to prepare a huge amount of index samples in advance for analysis, but also A technical problem arises in that accurate analysis results may not be obtained because information and the like are not taken into account. Moreover, the analysis process is also complicated, and a technical problem arises that the processing load becomes extremely large.

Examples of the problems to be solved by the present invention are as described above. An object of the present invention is to provide a shunt sound analysis apparatus capable of analyzing the shunt sound acquired from a subject and suitably assisting diagnosis of shunt stenosis.

A shunt sound analysis apparatus for solving the above problems includes an acquisition means for acquiring shunt sound information about a shunt sound from around a shunt formation site of a subject, and based on the shunt sound information acquired by the acquisition unit, and an output means for outputting information indicating the audible intermittent degree of the shunt sound.

A shunt sound analysis method for solving the above problems includes an acquisition step of acquiring shunt sound information about a shunt sound from the vicinity of a shunt formation site of a subject, and based on the shunt sound information acquired in the acquisition step, and an outputting step of outputting information indicating an audible intermittent degree of the shunt sound.

A computer program for solving the above problems includes an acquisition step of acquiring shunt sound information about a shunt sound from around a shunt formation site of a subject; and an output step of outputting information indicating the degree of audible intermittent sound.

A recording medium for solving the above problems stores the computer program described above.

1 is a block diagram showing the overall configuration of a shunt sound analysis device according to an embodiment; FIG. 1 is a spectrogram (No. 1) showing an example of a time-frequency waveform; 4 is a graph showing an example of volume analysis waveform; 2 is a spectrogram (part 2) showing an example of a time-frequency waveform; It is a graph which shows an example of a frequency centroid analysis waveform. It is a flowchart which shows operation|movement of a personal difference input part. 10 is a histogram showing the distribution of minimum values of shunt sounds accumulated from the past of the subject. 10 is a histogram showing the distribution of maximum values of shunt sounds accumulated from the past of the subject. 4 is a spectrogram showing an example of a time-frequency waveform of one heart beat interval in normal state; FIG. 10 is a graph showing an example of a frequency barycenter analysis waveform of one heartbeat interval in normal time; FIG. 4 is a spectrogram showing an example of a time-frequency waveform of one heart beat interval when stenosis occurs. FIG. 10 is a graph showing an example of a frequency barycenter analysis waveform of one heart beat interval when stenosis occurs; FIG. 7 is a graph showing an example of the amount of sound volume change in one heart beat interval in normal times; FIG. 10 is a graph showing an example of the amount of change in sound volume in one heartbeat interval when constriction occurs. FIG. FIG. 11 is a graph showing an example of the minimum sound volume value in one heart beat interval in normal times; FIG. FIG. 11 is a graph showing an example of the minimum sound volume value in one heartbeat interval when constriction occurs; FIG. FIG. 10 is a graph showing an example of the sound volume attenuation rate of one heart beat interval in normal state; FIG. FIG. 10 is a graph showing an example of volume attenuation rate in one heart beat interval when constriction occurs. FIG. 7 is a graph showing an example of a volume analysis waveform when distortion occurs; 7 is a graph showing an example of a volume analysis differential waveform when distortion occurs.

<1>
A shunt sound analysis apparatus according to the present embodiment includes an acquisition unit that acquires shunt sound information about a shunt sound from around a shunt formation site of a subject; an output means for outputting information indicating the degree of audible intermittent sound.

When the shunt sound analysis device according to the present embodiment operates, first, the acquisition means acquires shunt sound information about the shunt sound from around the shunt formation site of the person to be measured. Here, the "shunt sound" is the blood flow sound acquired around the shunt formation site for drawing blood out of the body, and is the sound synchronized with the pulse of the subject. The shunt sound may be acquired using various sensors, and the acquisition method is not particularly limited. "Shunting sound information" is information including various parameters related to the shunt sound, and includes, for example, changes over time such as volume and frequency.

When the shunt sound information is acquired, various analyzes are performed by the output means, and information indicating the degree of intermittent audible shunt sound is output. Here, "information indicating the degree of intermittence" does not simply mean information indicating the degree of interruption of the shunt sound, but includes various parameters that cause intermittent sensation, which is used as an index for diagnosing stenosis. It means composite information or information in which a plurality of parameters related to intermittent feeling are integrated. Note that the information indicating the degree of intermittence is output, for example, in a numerical state or in a state visualized by a graph or chart.

According to research conducted by the inventors of the present application, when no stenosis has occurred at the shunt formation site, the shunt sound tends to be loud and continuous, whereas when the stenosis progresses, the shunt sound decreases. As a result, it has been found that high-pitched sound components due to discontinuities in sound and turbulent flow at the constriction site appear singly or in combination. Therefore, by using the information indicating the intermittent degree of the shunt sound, it is possible to easily and accurately diagnose whether or not the stenosis has occurred at the shunt formation site, or the degree of the stenosis. Specifically, stenosis diagnosis can be easily performed without using echo diagnosis or the like that has been conventionally used for stenosis diagnosis. In addition, the information indicating the intermittent degree of the shunt sound enables quantitative diagnosis that is not affected by the skill or experience of the doctor who diagnoses the stenosis.

As described above, according to the shunt sound analysis device according to the present embodiment, it is possible to favorably support the stenosis diagnosis of the shunt formation site.

<2>
In one aspect of the shunt sound analysis device according to the present embodiment, the output means includes volume information derivation means for deriving volume information indicating a change in volume of the shunt sound over time based on the shunt sound information. .

According to this aspect, it is possible to output information indicating the intermittent degree of the shunt sound based on the volume information derived by the volume information deriving means. In particular, according to research conducted by the inventors of the present application, it has been found that the intermittent feeling of the shunt sound greatly depends on changes in the volume of the shunt sound over time. Therefore, by using the derived volume information, it becomes possible to output more appropriate information as information indicating the intermittent degree of the shunt sound.

<3>
In the aspect provided with the volume information deriving means described above, the output means may have extraction means for extracting one cycle volume information corresponding to one cycle of the shunt sound from the volume information.

In this case, once the volume information is derived, one cycle volume information corresponding to one cycle of the shunt sound (in other words, one cycle of pulsation of the subject) is extracted. Therefore, it is possible to suitably analyze the volume change in one cycle of the shunt sound (for example, the volume change from systole to diastole). Therefore, it is possible to output more appropriate information as information indicating the intermittent degree of the shunt sound.

<4>
In the aspect provided with the above-mentioned extraction means, the output means outputs the difference between the maximum and minimum values of the volume of the shunt sound in the one-cycle volume information and the difference between the maximum and minimum values of the shunt sound during normal operation. In contrast, it may have a first quantifying means for quantifying the degree of intermittence.

In this case, first, the minimum value (e.g. diastolic volume) is subtracted from the maximum value (e.g. systolic volume) of the shunt sound volume in the one-cycle volume information to obtain the volume change amount in one cycle. Then, the degree of intermittence is quantified by comparing the calculated amount of change in volume with the amount of change in volume in the normal state that has been stored in advance.

According to research conducted by the inventors of the present application, it has been found that when the intermittent degree of the shunt sound increases, the amount of volume change in one cycle tends to increase. Therefore, when the calculated volume change amount is larger than the normal volume change amount, a numerical value indicating a high degree of intermittence may be output. On the other hand, when the calculated volume change amount is smaller than the normal volume change amount, a numerical value indicating a low degree of intermittence may be output.

<5>
Alternatively, in a mode comprising an extracting means, the output means compares the minimum volume value of the shunt sound in the one-cycle volume information with the minimum value of the shunt sound in a normal state, and digitizes the intermittent degree. It may have a second digitizing means.

In this case, first, the minimum value of the volume of the shunt sound (for example, the volume at the end of diastole) in the one-cycle volume information is extracted. Then, the degree of intermittence is quantified by comparing the extracted minimum value with the pre-stored minimum value in the normal state.

According to research conducted by the inventor of the present application, it has been found that when the intermittent degree of the shunt sound increases, the minimum value of the sound volume in one cycle tends to decrease. Therefore, when the calculated minimum value is larger than the normal minimum value, a numerical value indicating a low degree of intermittence may be output. On the other hand, when the calculated minimum value is smaller than the normal minimum value, a numerical value with a high degree of intermittence may be output.

<6>
Alternatively, in a mode comprising an extraction means, the output means includes third quantification means for quantifying the intermittent degree based on the slope from the maximum value to the minimum value of the volume of the shunt sound in the one-cycle volume information. may have.

In this case, first, the slope from the maximum value to the minimum value of the volume of the shunt sound in the one-period volume information (in other words, the attenuation rate from the maximum value to the minimum value) is calculated. Then, the degree of intermittence is quantified based on the calculated inclination.

According to research conducted by the inventors of the present application, it has been found that when the intermittent degree of the shunt sound increases, there is a tendency for the sound volume to attenuate rapidly from the systole to the diastole. Therefore, the larger the calculated change in inclination, the higher the degree of intermittence should be output as a numerical value.

<7>
Alternatively, in a mode comprising an extracting means, the output means compares the difference between the maximum value and the minimum value of the volume of the shunt sound in the one-period volume information with the difference between the maximum value and the minimum value of the shunt sound in the normal state. and fourth digitizing means for digitizing the degree of intermittence.

In this case, first, the maximum value and the minimum value of the volume of the shunt sound in the one-cycle volume information are extracted. The maximum and minimum values can be extracted using, for example, a smoothed differential waveform. Subsequently, the minimum value is subtracted from the extracted maximum value, and the volume change amount from the maximum value to the minimum value is obtained. Then, the degree of intermittence is quantified by comparing the calculated amount of change in volume with the amount of change in volume in the normal state that has been stored in advance.

According to research conducted by the inventors of the present application, when the intermittent degree of the shunt sound increases, the amount of volume change from the maximum value to the minimum value increases. distortion) has been found. Therefore, when the calculated volume change amount is larger than the normal volume change amount, a numerical value indicating a high degree of intermittence may be output. On the other hand, when the calculated volume change amount is smaller than the normal volume change amount, a numerical value indicating a low degree of intermittence may be output.

<8>
In another aspect of the shunt sound analysis device according to the present embodiment, the output means derives distribution information indicating the volume of the shunt sound over time for each frequency based on the shunt sound information. have the means.

According to this aspect, it is possible to output information indicating the intermittent degree of the shunt sound based on the distribution information derived by the distribution information deriving means. In particular, according to research conducted by the inventors of the present application, it has been found that the intermittent feeling of the shunt sound greatly depends on the volume distribution of the shunt sound for each frequency. Therefore, by using the derived distribution information, it is possible to output more appropriate information as the information indicating the intermittent degree of the shunt sound.

<9>
In the aspect provided with the distribution information deriving means described above, the output means compares the amount of change in the center of gravity of the frequency over time indicated by the distribution information with the amount of change in the center of frequency over time under normal conditions, It may have a fifth quantifying means for quantifying the degree of intermittence.

In this case, first, the amount of change in the center of gravity of the frequency over time is calculated from the distribution information. The amount of change in the frequency center of gravity can be calculated, for example, by subtracting the minimum center-of-gravity frequency from the maximum center-of-gravity frequency. Then, the degree of intermittence is quantified by comparing the calculated change amount of the frequency center of gravity with the previously stored change amount of the frequency center of gravity during normal operation.

According to research conducted by the inventor of the present application, it has been found that when the intermittent degree of the shunt sound increases, the amount of change in the center of gravity of the frequency tends to increase (specifically, the amount of high-frequency components increases). Therefore, when the calculated amount of change in the frequency center of gravity is greater than the amount of change in the normal frequency center of gravity, a numerical value indicating a high degree of intermittence may be output. On the other hand, when the calculated amount of change in the center of gravity of frequency is smaller than the amount of change in the center of gravity of frequency under normal conditions, a numerical value indicating a low intermittent degree may be output.

<10>
In another aspect of the shunt sound analysis device according to the present embodiment, the output means has at least two types of the first to fifth digitizing means, and the first to fifth digitizing means Each of the intermittence degrees thus obtained is comprehensively determined, and information indicating the intermittence degree is output.

According to this aspect, since the degree of intermittence quantified using different indexes is integrally determined, it is possible to output information indicating a more accurate degree of intermittence. For the integrated determination, for example, the degree of intermittence quantified by each quantifying means may be normalized and then given a predetermined weight. The information indicating the degree of intermittence to be output may be, for example, one value that emphasizes hearing sensation, one value that emphasizes waveform intermittence, or an intermittent tendency that matches auditory sensation and a waveform-like intermittent tendency. and the two values shown.

<11>
A shunt sound analysis method according to the present embodiment includes an acquisition step of acquiring shunt sound information about a shunt sound from the vicinity of a shunt formation site of a subject; and an output step of outputting information indicating the degree of audible intermittent sound.

According to the shunt sound analysis method according to the present embodiment, similarly to the above-described shunt sound analysis apparatus according to the present embodiment, information indicating the audible intermittent degree of the shunt sound is output. Therefore, it is possible to favorably assist the stenosis diagnosis of the shunt formation site.

<12>
A computer program according to the present embodiment comprises an acquisition step of acquiring shunt sound information about a shunt sound from the vicinity of a shunt formation site of a subject; and an output step of outputting information indicating the degree of audible discontinuity.

According to the computer program according to this embodiment, it is possible to cause a computer to execute each step of the above-described shunt sound analysis method according to this embodiment. Therefore, it is possible to favorably assist the stenosis diagnosis of the shunt formation site.

<13>
A recording medium according to the present embodiment records the computer program described above.

According to the recording medium of the present embodiment, by executing the recorded computer program, it is possible to output information indicating the auditory intermittent degree of the shunt sound. Therefore, it is possible to favorably assist the stenosis diagnosis of the shunt formation site.

The effects and other gains of the shunt sound analysis device, 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 analysis device will be described in detail with reference to the drawings.

<Device configuration>
First, referring to FIG. 1, the overall configuration of the shunt sound analysis device according to the present embodiment will be described. FIG. 1 is a block diagram showing the overall configuration of the shunt sound analysis device according to the embodiment.

In FIG. 1, the shunt sound analysis apparatus according to this embodiment includes an audio signal input unit 110, an audio signal analysis unit 120, an individual difference input unit 130, a global feature extraction unit 140, and a local feature extraction unit 150. , an integrated determination unit 160 , and a display unit 170 .

The audio signal input unit 110 is composed of, for example, a vibration sensor or the like, and detects the shunt sound from the shunt formation site of the person to be measured. The shunt sound detected by the audio signal input section 110 is configured to be output to the audio signal analysis section 120 as an audio signal. The audio signal input unit 110 is a specific example of the “acquisition unit”.

The audio signal analysis unit 120 performs time-frequency analysis on the audio signal input from the audio signal input unit 110 . The analysis result by the audio signal analysis unit 120 is configured to be output to each of the global feature extraction unit 140 and the local feature extraction unit 150 .

The individual difference input unit 130 calculates a parameter indicating the shunt sound specific to the person to be measured based on the sound volume waveform obtained from the person to be measured in advance. The subject-specific parameters calculated by the individual difference input section 130 are configured to be output to each of the global feature extraction section 140 and the local feature extraction section 150 .

The global feature extraction unit 140 extracts global features (specifically, the amount of change in the center of gravity of the frequency, the amount of volume change in one heartbeat interval, and the minimum volume value in one heartbeat interval) among the factors that cause the intermittent feeling of the shunt sound. ) is extracted, and information indicating the degree of intermittence is calculated based on each feature. It should be noted that global feature extraction section 140 takes into account information specific to the subject input from individual difference input section 130 when extracting global features. Information indicating the degree of intermittence calculated by the global feature extraction unit 140 is configured to be output to the integrated determination unit 160 .

The local feature extraction unit 150 extracts local features (specifically, the volume attenuation rate of one heartbeat interval and the volume change distortion) among the factors that cause the intermittent feeling of the shunt sound, and extracts the to calculate information indicating the degree of intermittence. It should be noted that local feature extraction section 150 takes into account information specific to the subject input from individual difference input section 130 when extracting local features. Information indicating the degree of intermittence calculated by the local feature extraction unit 150 is configured to be output to the integrated determination unit 160 .

The integrated determination unit 160 comprehensively determines the information indicating the intermittent degree of the shunt sound input from the global feature extraction unit 140 and the local feature extraction unit 150, and calculates information indicating the possibility of stenosis at the shunt formation site. do. Information indicating the possibility of stenosis calculated by the integrated determination unit 160 is configured to be output to the display unit 170 . The integrated determination unit 160 is a specific example of the “output unit”.

The display unit 170 is configured, for example, as a liquid crystal display or the like, and visually displays the information indicating the possibility of stenosis (information indicating the intermittent degree of the shunt sound) output from the integrated determination unit 160 to a user such as a doctor. It is configured so that it can be presented to In addition, the display unit 170 displays any of the above-described amount of change in the center of gravity of the frequency, amount of volume change in one heartbeat interval, minimum volume value in one heartbeat interval, volume attenuation rate in one heartbeat interval, and volume change distortion. You may make it present visually.

<Description of operation>
Next, the operation of the shunt sound analysis device according to this embodiment will be described. In the following description, of the parts included in the shunt sound analysis device according to the present embodiment, the parts specific to this embodiment (that is, the audio signal analysis unit 120, the individual difference input unit 130, the global feature extraction unit 140, The operations of the local feature extraction unit 150 and the integrated determination unit 160) will be described in detail.

<Audio signal analysis part>
First, the operation of the audio signal analysis unit 120 will be specifically described with reference to FIGS. 2 to 5. FIG. FIG. 2 is a spectrogram (Part 1) showing an example of the time-frequency waveform, and FIG. 3 is a graph showing an example of the volume analysis waveform. FIG. 4 is a spectrogram (part 2) showing an example of a time-frequency waveform, and FIG. 5 is a graph showing an example of a frequency centroid analysis waveform.

In FIG. 2, the audio signal analysis unit 120 performs time-frequency analysis of the audio signal input from the audio signal input unit 110, and acquires a time-frequency waveform. The time-frequency waveform indicates the power of the shunt sound at each frequency in time series. The audio signal analysis unit 120 calculates the volume of the shunt sound from the time-frequency waveform. Specifically, the audio signal analysis unit 120 calculates the volume y(t) of the shunt sound using the following formula (1).

なお、ｆ（ｎ）は周波数であり、ｐ（ｎ）は周波数ｆ（ｎ）におけるパワーである。

Note that f(n) is the frequency and p(n) is the power at the frequency f(n).

As shown in FIG. 3, the volume analysis waveform (that is, the waveform representing the temporal change in the volume of the shunt sound) calculated by the audio signal analysis unit 120 is a periodic waveform corresponding to the pulsation of the subject. The volume analysis waveform is divided into each heartbeat interval, and then output to the global feature extraction section 140 and the local feature extraction section 150, which are used to extract global features and local features, respectively.

In FIG. 4, the audio signal analysis unit 120 further calculates the frequency centroid of the shunt sound from the time-frequency waveform. Specifically, the audio signal analysis unit 120 calculates the frequency center of gravity g(t) of the shunt sound using the following formula (2).

図５に示すように、音声信号解析部１２０が算出した周波数重心解析波形（即ち、シャント音の周波数重心の時間変化を示す波形）は、音量解析波形と同様に、被測定者の脈動に応じた周期的な波形となる。周波数重心解析波形は、大局的特徴抽出部１４０に出力され、大局的特徴（具体的には、周波数重心の変化）の抽出に利用される。

As shown in FIG. 5, the frequency-center-of-gravity analysis waveform (that is, the waveform representing the temporal change in the frequency-center of gravity of the shunt sound) calculated by the audio signal analysis unit 120 corresponds to the pulsation of the subject, similarly to the volume analysis waveform. It becomes a periodic waveform. The frequency centroid analysis waveform is output to global feature extraction section 140 and used to extract global features (specifically, changes in frequency centroid).

<Individual difference input section>
Next, the operation of the individual difference input unit 130 will be specifically described with reference to FIGS. 6 to 8. FIG. FIG. 6 is a flow chart showing the operation of the individual difference input unit 130. As shown in FIG. FIG. 7 is a histogram showing the distribution of the minimum values of the shunt sounds accumulated from the past of the subject, and FIG. 8 is a histogram showing the distribution of the maximum values of the shunt sounds accumulated from the past of the subject. is.

In FIG. 6, an individual difference input unit 130 calculates a parameter specific to the person to be measured from an audio signal obtained in advance. Specifically, when the individual difference input unit 130 acquires an audio signal (step S101), a volume analysis waveform is acquired by volume analysis (step S102). The volume analysis waveform is divided by one heartbeat interval (step S103), and the maximum value and minimum value in each interval are detected (step S104). The detected maximum and minimum values are stored in memory (step S105), and the most frequent values from the histogram are determined as the maximum and minimum volume values of the shunt sound specific to the subject. (step S106).

In FIGS. 7 and 8, when the frequency of detection of the minimum value and the maximum value is distributed as shown in the figures, the volume minimum value of the shunt sound peculiar to the person to be measured is determined to be approximately 137. FIG. Similarly, the volume maximum value of the subject's unique shunt sound is determined to be approximately 425.

The measured person's unique maximum volume value and minimum volume value determined in this way are used to correct individual differences in the volume of the shunt sound obtained as the volume analysis waveform. Corrected volume Y(t) is obtained by taking the average value of volume y(t) obtained by volume analysis as y_ave, the maximum value as y_max, and the minimum value as y_min. Assuming that the value is Y_min, it can be calculated using the following formula (3).

Y(t)=y_ave+(y(t)-y_ave)*(y_max-y_min)/(Y_max-Y_min) (3)
Although the description here is omitted, the maximum value and minimum value specific to each subject may also be determined for the center of gravity of the frequency. Correction of individual differences in the center of gravity of frequency can also be performed in the same manner as the above-described correction of individual differences in sound volume.

<Global Feature Extraction Unit>
Next, the operation of the global feature extraction unit 140 will be specifically described with reference to FIGS. 9 to 16. FIG. A method of calculating the degree of intermittence for each of a plurality of features that can be extracted by the global feature extraction unit 140 will be described below.

<Amount of change in frequency center of gravity in one heartbeat interval>
First, with reference to FIGS. 9 to 12, a method of calculating the degree of discontinuity based on the amount of change in the center of gravity of the frequency in one heart beat interval will be described. Here, FIG. 9 is a spectrogram showing an example of the time-frequency waveform of one heartbeat interval in the normal state, and FIG. 10 is a graph showing an example of the frequency centroid analysis waveform of one heartbeat interval in the normal state. FIG. 11 is a spectrogram showing an example of a time-frequency waveform of one heartbeat interval when stenosis occurs, and FIG. 12 is a graph showing an example of a frequency barycentric analysis waveform of one heartbeat interval when stenosis occurs.

In FIGS. 9 and 10, the normal shunt sound is detected as sound containing many low frequencies. As can be seen from FIG. 10, the minimum value of the center of gravity of the frequency of the shunt sound under normal conditions is approximately 610 Hz, and the maximum value thereof is approximately 700 Hz. Therefore, the amount of change in the center of gravity of the frequency of the shunt sound (that is, maximum value-minimum value) is approximately 90 Hz in the normal state.

In FIGS. 11 and 12, the shunt sound at the time of stenosis (that is, the feeling of intermittent feeling is strong) is detected as a sound containing more high-frequency components than in the normal state. As can be seen from FIG. 12, the minimum value of the center of gravity of the frequency of the shunt sound when constriction occurs is about 590 Hz, and the maximum value is about 740 Hz. Therefore, the amount of change in the center of gravity of the frequency of the shunt sound (that is, maximum value-minimum value) at the time of constriction is approximately 150 Hz.

As can be seen from the above results, there is a clear difference between the normal shunt sound and the stenosis-induced shunt sound in the amount of change in the center of gravity of the frequency in one heart beat interval. Therefore, by comparing the amount of change in the frequency center of gravity obtained by the analysis of the frequency center of gravity with the amount of change in the frequency center of gravity under normal conditions, the possibility of occurrence of stenosis (that is, the intermittent degree of the shunt sound) can be calculated favorably.

<Volume change in one heartbeat interval>
Next, with reference to FIGS. 13 and 14, a method of calculating the degree of intermittence based on the amount of volume change in one heartbeat interval will be described. Here, FIG. 13 is a graph showing an example of the amount of sound volume change in one heartbeat interval in normal times. Also, FIG. 14 is a graph showing an example of the amount of volume change in one heartbeat interval when constriction occurs.

In FIG. 13, the normal shunt sound has a small difference between the maximum and minimum volume values in one heart beat interval. That is, the normal shunt sound has a relatively small change in volume from systole to diastole.

In FIG. 14, the difference between the maximum value and the minimum value of the sound volume in one heart beat interval is larger in the shunt sound at the time of stenosis than in the normal time. That is, the shunt sound at the time of stenosis has a relatively large change in volume from systole to diastole.

Therefore, the possibility of occurrence of stenosis (that is, the intermittent degree of the shunt sound) depends on how large the volume change amount in one heartbeat interval obtained from the volume analysis waveform is compared to the volume change amount in one heartbeat interval under normal conditions. can be suitably calculated.

<Minimum volume value for one heartbeat interval>
Next, with reference to FIGS. 15 and 16, a method of calculating the degree of discontinuity based on the minimum sound volume value of one heartbeat interval will be described. Here, FIG. 15 is a graph showing an example of the minimum sound volume value in one heart beat interval in the normal state. Also, FIG. FIG. 11 is a graph showing an example of the minimum sound volume value in one heartbeat interval when constriction occurs; FIG.

In FIG. 15, the normal shunt sound has a relatively large minimum volume value in one heart beat interval. Specifically, the normal shunt sound is maintained at a relatively high volume even during diastole when the volume decreases.

In FIG. 16, the shunt sound at the time of stenosis has a larger minimum volume value in one heartbeat interval than at normal time. Specifically, the volume of the shunt sound at the time of stenosis increases to a value close to the normal volume during systole, but the volume during diastole is greatly attenuated.

Therefore, the possibility of occurrence of stenosis (that is, the intermittent degree of shunt sound) depends on how small the minimum volume value in one heartbeat interval obtained from the volume analysis waveform is compared to the minimum volume value in one heartbeat interval under normal conditions. can be suitably calculated.

<Local Feature Extractor>
Next, with reference to FIGS. 17 to 20, the operation of local feature extraction section 150 will be specifically described. A method of calculating the degree of intermittence for each of a plurality of features that can be extracted by the local feature extraction unit 150 will be described below.

<Volume attenuation rate for one heartbeat interval>
First, with reference to FIGS. 17 and 18, a method of calculating the degree of intermittence based on the volume attenuation rate of one heartbeat interval will be described. Here, FIG. 17 is a graph showing an example of the volume attenuation rate of one heart beat interval in the normal state. Also, FIG. 18 is a graph showing an example of the sound volume attenuation rate in one heart beat interval when constriction occurs.

In FIG. 17, the normal shunt sound shows a linear attenuation of the volume in one heart beat interval. Specifically, in the normal shunt sound, the slope (attenuation rate) from the maximum value to the minimum value of the volume is constant.

In FIG. 18, in the shunt sound at the time of stenosis, there is a portion where the sound volume is abruptly attenuated in one heart beat interval. Specifically, the volume of the shunt sound at the onset of stenosis decreases significantly from the end of systole to the beginning of diastole, and then decreases relatively gently. Therefore, the slope (attenuation rate) from the maximum value to the minimum value of the volume is not constant.

Therefore, the probability of occurrence of constriction (that is, the intermittent degree of shunt sound) can be preferably calculated by using the volume attenuation rate in one heart beat interval obtained from the volume analysis waveform. The attenuation factor dr can be calculated using the following formula (4).

なお、Ｎは１心拍区間のデータ数であり、ｍａｘ及びｍｉｎは、それぞれ１心拍区間内の音量の最大値及び最小値である。減衰率ｄｒは、図１８中に破線で示す直線的な減衰率である０．５を基準として、そこからの乖離が大きいほど小さい値として算出される。よって、算出された減衰率ｄｒが０．５よりどの程度小さいかによって、狭窄発生可能性（即ち、シャント音の断続度合い）を好適に算出できる。

Note that N is the number of data in one heartbeat interval, and max and min are the maximum and minimum volume values in one heartbeat interval, respectively. The attenuation rate dr is calculated as a smaller value as the deviation from 0.5, which is the linear attenuation rate indicated by the dashed line in FIG. 18, becomes larger. Therefore, the probability of occurrence of constriction (that is, the intermittent degree of the shunt sound) can be preferably calculated depending on how much the calculated attenuation rate dr is smaller than 0.5.

<Distortion of volume change>
Next, with reference to FIGS. 19 and 20, a method of calculating the degree of discontinuity based on the distortion of volume change that occurs from systole to diastole will be described. FIG. 19 is a graph showing an example of volume analysis waveforms when distortion occurs. FIG. 20 is a graph showing an example of a volume analysis differential waveform when distortion occurs.

In FIG. 19, the shunt sound at the time of stenosis may be distorted in changes in volume from the end of systole to the beginning of diastole. Specifically, as shown in the portion surrounded by the dashed line in the figure, the volume may rise temporarily after sharply dropping. Therefore, it is possible to calculate the possibility of constriction occurrence (that is, the intermittent degree of the shunt sound) from the existence of the maximum value in the volume analysis waveform.

In FIG. 20, the maximum value in the volume analysis waveform can be easily detected by using the smoothed differentiated waveform of the volume analysis waveform. Specifically, it can be preferably detected by calculating the difference between the maximum value and the minimum value of the differentiated waveform of the volume analysis waveform.

<Integrated judgment part>
Finally, the operation of the integrated determination unit 160 will be specifically described. In the following, ρ ₁ is the feature amount extracted as the amount of change in the frequency center of gravity, which is a global feature, ρ 2 is the feature amount extracted as the amount of volume change in one heartbeat interval, and ρ ₂ is the feature amount extracted as the amount of volume change in one heartbeat interval. ρ ₃ , ρ ₄ , and ρ ₅ , respectively. Each direct quantity ρ ₁ to ρ ₅ is normalized (for example, it is adjusted to 0 when the intermittent degree is the strongest and 100 when the intermittent degree is the weakest (that is, normal)). shall be

The integrated determination unit 160 weights each of the feature quantities ρ ₁ to ρ ₅ to calculate the degree of intermittence. Specifically, the degree of intermittence is calculated using the following formula (5), with weights ω ₁ to ω ₅ corresponding to the respective feature quantities ρ ₁ to ρ ₅ .

Intermittence= _ω1 × _ρ1 + _ω2 × _ρ2 + _ω3 × _ρ3 + _ω4 × _ρ4 + _ω5 × _ρ5 (5)
Here, when the intermittent degree is output as a single numerical value that emphasizes fitting to the sense of hearing, the weights ω ₁ to ω ₅ should be set so that the relationships ω1, ω2, ω3 >> ω4, ω5 are established. . The degree of intermittence obtained in this way is suitable, for example, for determining whether or not dialysis on the day is possible at the site of artificial dialysis.

Further, when the discontinuity is output as one numerical value that emphasizes the discontinuity of the waveform, the weights ω ₁ to ω ₅ should satisfy the relationships ω4, ω5>ω2>ω1, ω3. The intermittence obtained in this way is suitable for comparison with echodiagnosis, for example.

Alternatively, when outputting the degree of discontinuity as two values, a numerical value emphasizing matching the audibility and a numerical value emphasizing the discontinuity of the waveform, the values are calculated separately using the following formulas (6) and (7). do it.

Intermittent degree by hearing = ω ₁ × ρ ₁ + ω ₂ × ρ ₂ + ω ₃ × ρ ₃ (6)
Intermittence by waveform=ω ₄ ×ρ ₄ +ω ₅ ×ρ ₅ (7)
The degree of intermittence obtained in this way is suitable, for example, when judging the degree of progress of intermittence (that is, stenosis) from changes over time in both auditory sense and waveform.

It should be noted that each of the feature quantities ρ ₁ to ρ ₅ may be output independently without performing integrated determination by the integrated determination unit 160 . In this case, the display unit 170 displays, for example, numerical values corresponding to the feature amounts ρ ₁ to ρ ₅ separately or as a radar chart. As a result, a user, such as a doctor, can diagnose stenosis from a plurality of viewpoints using each of the feature quantities ρ ₁ to ρ ₅ . Alternatively, using each of the feature values ρ ₁ to ρ ₅ , a doctor or the like can make an integrated judgment by himself/herself to diagnose stenosis.

As described above, according to the shunt sound analysis device according to the present embodiment, information indicating the degree of intermittence is output as the analysis result of the shunt sound. Therefore, it is possible to favorably assist in stenosis diagnosis at the shunt formation site.

The present invention is not limited to the above-described embodiments, and can be modified as appropriate within the scope not contrary to the gist or idea of the invention that can be read from the scope of claims and the entire specification. The apparatus, shunt sound analysis method, computer program and recording medium are also included in the technical scope of the present invention.

110 audio signal input unit 120 audio signal analysis unit 130 individual difference input unit 140 global feature extraction unit 150 local feature extraction unit 160 integration determination unit 170 display unit

Claims (10)

Acquisition means for acquiring shunt sound information about the shunt sound from around the shunt formation site of the subject;
extracting means for extracting one-cycle volume information corresponding to one cycle of the shunt sound from the shunt sound information;
(i) volume change amount, (ii) minimum volume, (iii) volume attenuation rate, (iv) volume change distortion, and (v) frequency center of gravity of the one cycle derived based on the one cycle volume information and output means for outputting information indicating the possibility of constriction of the shunt formation site, which is generated based on any of the amount of change in the shunt sound analysis apparatus. The volume change amount is information indicating the difference between the maximum value and the minimum value of the volume of the shunt sound in the one-period volume information,
The output means has a first quantifying means for quantifying the possibility of constriction by comparing the volume change amount and the difference between the maximum value and the minimum value of the shunt sound in a normal state. Item 2. The shunt sound analysis device according to item 1. The output means has a second quantifying means for quantifying the possibility of constriction by comparing the minimum value of the volume of the shunt sound in the one-cycle volume information with the minimum value of the shunt sound in a normal state. The shunt sound analysis device according to claim 1, characterized in that: The volume attenuation rate is information indicating the slope from the maximum value to the minimum value of the volume of the shunt sound in the one-cycle volume information,
2. The shunt sound analysis apparatus according to claim 1, wherein said output means has third quantifying means for quantifying said constriction possibility based on said volume attenuation rate. The volume change distortion is information indicating the difference between the maximum value and the minimum value of the volume of the shunt sound in the one-cycle volume information,
The output means has a fourth quantifying means for quantifying the possibility of constriction by comparing the distortion of the volume change with the difference between the maximum value and the minimum value of the shunt sound in the normal state. The shunt sound analysis device according to claim 1. further comprising distribution information deriving means for deriving distribution information indicating the volume of the shunt sound over time for each frequency based on the one-cycle volume information;
The amount of change in the frequency center of gravity is information indicating the amount of change in the frequency center of gravity over time indicated by the distribution information,
The output means has a fifth quantifying means for quantifying the possibility of stenosis by comparing the amount of change in the frequency center of gravity with the amount of change in the frequency center of gravity over time in a normal state. The shunt sound analysis device according to claim 1. The output means is
(ii) minimum volume; (iii) volume attenuation rate; (iv) volume change distortion;
Integratively determining the derived at least two types and outputting information indicating the possibility of stenosis
The shunt sound analysis device according to claim 1, characterized in that: an acquisition step of acquiring shunt sound information about the shunt sound from around the shunt formation site of the subject;
an extracting step of extracting one-cycle volume information corresponding to one cycle of the shunt sound from the shunt sound information;
(i) volume change amount, (ii) minimum volume, (iii) volume attenuation rate, (iv) volume change distortion, and (v) frequency of the one cycle derived based on the one cycle volume information an output step of outputting information indicating the possibility of stenosis of the shunt formation site generated based on any of the amount of change in the center of gravity;
A shunt sound analysis method comprising: an acquiring step of acquiring shunt sound information about the shunt sound from around the shunt formation site of the subject;
an extracting step of extracting one-cycle volume information corresponding to one cycle of the shunt sound from the shunt sound information;
(i) volume change amount, (ii) minimum volume, (iii) volume attenuation rate, (iv) volume change distortion, and (v) frequency of the one cycle derived based on the one cycle volume information an output step of outputting information indicating the possibility of stenosis of the shunt formation site generated based on any of the amount of change in the center of gravity;
A computer program characterized by causing a computer to execute A recording medium on which the computer program according to claim 9 is recorded.

Images