JP6036242B2 - Heart sound information processing apparatus, heart sound information processing method and program - Google Patents

Description

  The present invention relates to a heart sound information processing apparatus, a heart sound information processing method, and a program that analyze a heart sound and appropriately detect various characteristic portions from the heart sound.

  Various information is included in the heart sound, and the doctor diagnoses the presence and type of abnormality by listening to the heart sound using a stethoscope. However, diagnosis using a stethoscope requires special skills and training, and it is extremely difficult for ordinary people to perform (for example, Non-Patent Document 1). Therefore, if it is possible to analyze heart sounds using a computer and determine the presence or absence and type of abnormality, it is considered that this can be used to expand home medical care.

  In Patent Document 1, a sound signal from the heart is recorded to detect a peak position, and the peak position is classified into various heart sounds and heart noises by cluster analysis, and the types of heart sounds and heart noises included in the sound signals are disclosed. A system for determining the presence / absence and type of abnormality from the combination of the above is disclosed. Patent Document 2 discloses a technique for visualizing detected heart sounds.

JP-T-2004-529720 Special table 2009-535106 gazette

  Heart sounds detected in the actual environment are affected by noise from the inside and outside of the body, and even if the heart sounds containing these noises are subjected to frequency analysis as they are, correct heart sound detection results cannot be obtained. A person who does not have expertise in the field cannot perform a filter operation to cut noise and cannot obtain correct information.

  An object of the present invention is to provide a heart sound information processing apparatus, a heart sound information processing method, and a program for appropriately detecting various characteristic portions in a heart sound from a heart sound signal including noise.

  In order to achieve the above object, a heart sound information processing apparatus (100) according to the present invention includes a heart noise analysis signal generation unit (210) that generates a heart noise analysis signal for analyzing heart noise constituting a heart sound, A cardiac noise detection unit (310) that detects cardiac noise based on the cardiac noise analysis signal generated by the cardiac noise analysis signal generation unit (210), and the cardiac noise detection unit (310) uses the cardiac noise analysis signal. A heart noise sounding interval detection unit (311) for detecting a heart noise sounding interval, and a heart noise outline for determining whether or not the heart noise analysis signal is a heart noise based on a rough shape of the heart noise in the heart noise sounding interval. A shape determination unit (312) is provided.

  The heart sound information processing method according to the present invention includes a heart noise analysis signal generation step for generating a heart noise analysis signal for analyzing heart noise constituting heart sounds, and the heart noise analysis signal generation step. A heart noise sounding interval detection step for detecting a heart noise sounding interval using a noise analysis signal, and the heart noise analysis signal is converted into a heart noise according to an outline of the heart noise detected in the heart noise sounding interval detection step. A cardiac noise outline determining step for determining whether or not

  Further, a program according to the present invention includes a heart noise analysis signal generation step for generating a heart noise analysis signal for analyzing heart noise constituting heart sounds in a computer (120) included in the heart noise information processing apparatus (100). A heart noise sounding interval detection step for detecting a heart noise sounding interval using the heart noise analysis signal generated in the heart noise analysis signal generation step, and a heart in the heart noise sounding interval detected in the heart noise sounding interval detection step. A cardiac noise outline determining step for determining whether or not the cardiac noise analysis signal is a cardiac noise based on the outline of the noise is performed.

  ADVANTAGE OF THE INVENTION According to this invention, the various characteristic location in a heart sound can be detected appropriately from the heart sound signal containing noise.

It is a block diagram of the configuration of the heart sound information processing apparatus. It is a block diagram of the configuration of a cardiac noise detector. It is a flowchart which shows the example of a cardiac noise detection process. It is the conceptual diagram which showed the example which acquires an approximated curve from a heart sound signal, and also acquires peak area information (binarization signal). It is the conceptual diagram which showed the example which acquires an approximated curve from a heart sound signal, and also acquires peak area information (binarization signal). It is a graph which shows the frequency component of heart noise. It is a graph which shows the frequency component of the noise which is not a heart noise. It is a flowchart which shows the example of a cardiac noise outline determination process. It is the figure which represented the outline of the cardiac noise area with the approximated curve. It is the figure which represented the outline of the area which is not a heart noise with the approximated curve. It is a conceptual diagram showing the determination by the frame which normalized the outline of the heart noise. It is a figure which shows the example which normalized the heart noise sounding area. It is a figure which shows the example which normalized the heart noise sounding area. It is a figure which shows the example of the cardiac noise determination by normalization information. It is a figure which shows the example of the cardiac noise determination by normalization information. It is a figure which shows the example of the cardiac noise determination by normalization information. It is an example of the display information which modeled the heart sound signal containing a heart noise. It is an example of a display screen. It is a block diagram of the III sound detector. It is a flowchart which shows the example of a III sound detection process. It is a figure which shows the example of the peak position of a heart sound. It is a figure which shows the example of the peak position of a heart sound. It is a figure which shows the example for 1 period of the peak of a heart sound. It is a figure which shows the example for 1 period of the peak of a heart sound. It is a figure which shows the example of a clustering result. It is a figure which shows the example of a clustering result. It is the figure which represented the signal of acute myocardial infarction by the III sound analysis signal. It is an example of the display information which modeled the heart sound signal containing III sound.

  Hereinafter, a heart sound information processing apparatus 100 according to an embodiment of the present invention will be described with reference to the drawings. The heart sound information processing apparatus 100 may have any form such as a general-purpose PC (Personal Computer), various information terminals, a dedicated device, and the like, by running a program executable by a computer included in the various apparatuses. The information processing apparatus 100 may be realized.

  FIG. 1 is a block diagram showing components of the heart sound information processing apparatus 100. The heart sound information processing apparatus 100 includes a signal input unit 110, an arithmetic processing unit 120, a display unit 130, and a storage unit 140.

  The signal input unit 110 is an interface that receives an input of a heart sound signal, and includes, for example, a voice input terminal, an amplification device, and the like. The input heart sound signal may be a heart sound signal acquired by a microphone or the like, or may be a recorded heart sound signal. The input heart sound signal may be an analog signal or digital data. When the input heart sound signal is an analog signal, the signal input unit 110 may include an A / D (Analog to Digital) converter. When the input heart sound signal is recorded data and stored in the storage unit 140, the signal input unit 110 accepts an input of the heart sound signal in the storage unit 140.

  The arithmetic processing unit 120 includes a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), a DSP (Digital Signal Processor), and the like, and executes various programs stored in the ROM. Thus, operation control of each part constituting the heart sound information processing apparatus 100, processing or calculation of signals and data input from each part, file processing, and the like are performed.

  The display unit 130 includes, for example, a liquid crystal display element, an organic EL (Electro Luminescence) display element, and a circuit unit that drives them, and displays various display contents and display forms under the control of the arithmetic processing unit 120.

  The storage unit 140 includes, for example, various memories and an HDD (Hard Disk Drive), and is generated by the arithmetic processing unit 120, various data necessary for the heart sound information processing apparatus 100, heart sound signal data input from the signal input unit 110, and so on. Alternatively, various detected data are stored, and a storage operation and a read operation are performed by the processing of the arithmetic processing unit 120. The storage unit 140 is not limited to the one built in the heart sound information processing apparatus 100 but may be an external storage device connected by a predetermined interface. Examples of the external storage device include a USB memory connected to a USB (Universal Serial Bus) terminal, an external HDD device, and a memory card connected by a predetermined memory card slot. The storage unit 140 may store the recorded heart sound signal.

  The arithmetic processing unit 120 implements an analysis signal generation unit 200, a heart sound detection unit 300, and a display information generation unit 400 by executing a program.

  The analysis signal generation unit 200 performs processing for generating analysis signals of various elements constituting the heart sound. As a specific example, a cardiac noise analysis signal generation unit 210, an I sound II sound analysis signal generation unit 220, and a III sound analysis signal generation unit 230 are provided. Each element constituting the analysis signal generation unit 200 may be appropriately provided based on the heart sound to be analyzed. Besides these, the IV sound analysis signal generation unit, the diastolic noise analysis signal generation unit, and the click sound analysis A signal generation unit or the like may be provided.

  The heart noise analysis signal generator 210 generates an analysis signal for analyzing heart noise included in the heart sound. Similarly, the I sound II sound analysis signal generation unit 220 generates an analysis signal for analyzing the I sound and the II sound included in the heart sound, and the III sound analysis signal generation unit 230 analyzes the III included in the heart sound. An analysis signal for generating

  As an example of a specific method for generating each analysis signal, an input heart sound signal is divided at a predetermined time, converted into a frequency distribution by FFT processing or the like for each divided frame, and a predetermined frequency of each frame It is generated by calculating the sum of the band components. Alternatively, the method is not limited, such as filtering an input heart sound signal using a bandpass filter having a predetermined pass band, further converting the signal into a Hilbert transform, and using the envelope as an analysis signal.

  In addition, each analysis signal is generated not only from the input heart sound signal, but also based on the analysis signal generated in any element constituting the analysis signal generation unit 200. Good. For example, autocorrelation value calculation, smoothing process, downsampling, upsampling, etc. are performed based on any analysis signal.

  The heart sound detection unit 300 performs processing for analyzing various characteristic sounds included in the heart sound based on the various analysis signals generated by the analysis signal generation unit 200 and the heart sound signal input from the signal input unit 110. As a specific example, a cardiac noise detection unit 310, an I sound II sound detection unit 320, and a III sound detection unit 330 are provided. The elements constituting the heart sound detection unit 300 may be appropriately provided based on the heart sound to be analyzed, and in addition to these, an IV sound detection unit, a diastolic noise detection unit, a click sound detection unit, and the like are provided. Also good.

  Each component constituting analysis signal generation unit 200 and heart sound detection unit 300 may be incorporated in a modular manner according to the purpose of heart sound analysis. For example, a cardiac noise analysis module 410 including a cardiac noise analysis signal generation unit 210 and a cardiac noise detection unit 310 is used for cardiac noise analysis purposes. Further, the I sound II sound analysis module 420 including the I sound II sound analysis signal generation unit 220 and the I sound II sound detection unit 320 is used for the purpose of the I sound II sound analysis. Similarly, the III sound analysis module 430 including the III sound analysis signal generation unit 230 and the III sound detection unit 330 is used when detecting the III sound. Besides these, an IV sound analysis module, a click sound analysis module, a diastolic noise analysis module, an aortic regurgitation analysis module, a mitral regurgitation analysis module, and the like can be arbitrarily configured. In addition, each module may be a module configured in combination with hardware, or may be a module configured as software using a portable recording medium.

  FIG. 2 is a block diagram showing a configuration of the cardiac noise detection unit 310. The cardiac noise detection unit 310 includes a cardiac noise sound generation interval detection unit 311 and a cardiac noise outline that are configured as functions executed by the arithmetic processing unit 120. A determination unit 312 is provided.

  The heart noise sounding interval detector 311 detects a heart noise sounding interval using the heart noise analysis signal generated by the heart noise analysis signal generator 210.

  The cardiac noise outline determining unit 312 determines whether or not the cardiac noise analysis signal is cardiac noise based on the outline of the cardiac noise in the cardiac noise sounding section detected by the cardiac noise sounding section detecting unit 311.

  Next, heart noise detection processing for detecting heart noise will be described with reference to FIGS. The basic processing of various heart sounds analysis in the heart sound information processing apparatus 100 includes an analysis signal detection process for generating an analysis signal corresponding to the heart sound to be analyzed in the analysis signal generation unit 200, and the heart sound detection unit 300 based on the analysis signal. A heart sound detection process for detecting a heart sound is performed at. The same applies to the detection of cardiac noise. The cardiac noise analysis signal generation unit 210 generates a cardiac noise analysis signal, and the cardiac noise detection unit 310 detects cardiac noise based on the cardiac noise analysis signal. When the processing of FIG. 3 is executed, the heart sound information processing apparatus 100 can be mounted or executed by the heart noise analysis module 410.

  In the flowchart of FIG. 3, step S100 corresponds to the analysis signal detection process, and steps S110 to S120 correspond to the heart sound detection process.

  First, the heart noise analysis signal generation unit 210 generates a heart noise analysis signal from the heart sound signal input from the signal input unit 110 (step S100).

  The cardiac noise analysis signal generation processing in step S100 is based on the technology filed by the present applicant as Japanese Patent Application No. 2012-160252. First, the analysis signal generation unit 200 generates a signal indicating characteristics due to a time change for each frequency band of the heart sound signal input to the signal input unit 110. For example, processing such as short-time FFT is performed on the input heart sound signal to calculate a time frequency component, and a sum of frequency components in a predetermined frequency range at each time is calculated. The analysis signal generation unit 200 generates at least two signals, for example, a signal including any frequency component of 0 to 200 Hz and a signal including any frequency component of 100 to 2000 Hz. The former is a signal mainly showing characteristics due to temporal changes in heart sounds, and the latter is a signal mainly showing characteristics due to temporal changes in heart noise. Not limited to these two signals, three or more signals may be generated, and the frequency components of each signal may be exclusive or overlapped.

  Next, the analysis signal generation unit 200 approximates each signal obtained so far with a curve. For example, in the case of functioning as the cardiac noise analysis signal generation unit 210, an approximate curve is obtained for the signal indicating the characteristics due to the temporal change of the cardiac noise described above. As the approximate curve obtained here, it is preferable to obtain a plurality of approximate curves from one signal.

  4 and 5 are conceptual diagrams showing processing from obtaining an approximate curve from a heart sound signal to obtaining peak section information. FIG. 4A and FIG. 5A are signals that are the basis before obtaining the approximate curve. The result of the approximate curve varies depending on the difference in parameters such as the order of the curve used for the approximation and the interval used for the approximation. The solid lines in FIGS. 4B and 5B indicate the approximate curves of FIGS. 4A and 5A. 4B shows an approximate curve using a high-order curve, and FIG. 5B shows an approximate curve using a low-order curve. Even if the order of the approximate curve is the same, when the interval for obtaining the approximate curve is narrowed, the approximation is performed as shown in FIG. 4B, and when the interval is widened, the approximation is performed as shown in FIG. 5B. As described above, the approximate curve in FIG. 4B is an approximate curve close to the original signal, and the approximate curve in FIG. 5B well represents a rough tendency of the original signal.

  It should be noted that it is preferable to use at least two values for the interval for obtaining the approximate curve, one from the range of 50 to 200 ms and one from the range of 200 to 500 ms. The number of sections and approximate curves may be different for each signal. Each approximate curve obtained in this way is used as an analysis signal that is an output of the analysis signal generation unit 200, and in particular, the output of the cardiac noise analysis signal generation unit 210 based on a signal indicating characteristics due to temporal changes in cardiac noise is used as a cardiac noise analysis signal. To do.

  Next, based on the cardiac noise analysis signal generated in step S100, the cardiac noise sounding interval detector 311 detects a cardiac noise sounding interval that is a target for detecting cardiac noise (step S110).

  In the heart noise sounding section detection process in step S110, first, a second-order differential value is calculated for the approximate curve that is the heart-noise analysis signal generated in step S100, and the second-order differential value is a value that is greater than zero ( By binarizing into a section (convex downward) and a section that is 0 or less (upward convex), a section that is a target for detecting cardiac noise and another section are identified.

  4B and 5B show examples of binarization with respect to the example of the cardiac noise analysis signal. The lower dotted line in the rectangular waveform in the figure is a section where the second order differential value is greater than 0, and the upper one-dot chain line is a section where the second order differential value is 0 or less.

  Next, based on the peak section information indicated by the alternate long and short dash line in the binarized rectangular waveform shown in FIGS. 4B and 5B, whether or not cardiac noise is included in the peak section. An evaluation value used for the determination is calculated. First, the length of the peak section duration (hereinafter referred to as the peak duration) represented by the length of each of the alternate long and short dash lines in FIGS. 4B and 5B is calculated. Next, for each peak section information, the sum of the calculated peak duration sections is SUM, and the time length of the cardiac noise analysis signal is LEN.

  Next, the ratio RATIO (= SUM / LEN) of SUM and LEN, the average value AVE of the peak duration, and the standard deviation SD of the peak duration are calculated for each peak section information. Further, rSUM, rAVE, and rSD are respectively calculated as the ratio of SUM, AVE, and SD of each peak section information.

  Here, for example, it is assumed that SUM (m) and SUM (n) are calculated from arbitrary peak section information PEAK (m) and PEAK (n), respectively. In this case, rSUM is a value obtained by rSUM (m, n) = SUM (m) / SUM (n). Further, rAVE and rSD are also obtained in the same manner as rSUM. Further, the correlation value CORR of each peak section information may be obtained.

  RATIO, AVE, SD of each peak section information obtained in this way, and rSUM, rAVE, rSD, CORR by a combination of each peak section information are set as evaluation values.

  Next, the presence / absence of cardiac noise is determined based on the calculated evaluation value. Several methods for determining the presence or absence of cardiac noise based on the tendency of the evaluation value are conceivable. Here, the following method is exemplified.

  As an example of determination of the presence or absence of cardiac noise, each evaluation value is checked in order, and if there is a feature including cardiac noise, the determination value is counted up, and if not, the determination value is not counted up For each evaluation value. For example, when RATIO is larger than a predetermined value, 1 is added to the determination value, and when RATIO is smaller, it is not added (0 is added). By such processing, the determination value is high for data including cardiac noise, and the determination value is low for data not including cardiac noise. That is, the presence or absence of cardiac noise can be determined based on this determination value. If it can be determined that there is heart noise, a heart noise sounding section is detected based on peak section information generated by binarizing the analysis signal.

  Next, the cardiac noise outline determining unit 312 performs a cardiac noise outline determining process on the cardiac noise sounding section detected in step S110 (step S120). In the heart noise outline determination process in step S120, it is determined whether or not the frequency component of the heart noise sounding section is a rough shape that is heart noise, and a heart noise sounding section that is heart noise is extracted. It is processing.

  The outline of the heart noise used in the heart noise outline determining unit 312 will be described with reference to FIGS. FIG. 6 shows frequency components of heart noise, and FIG. 7 shows frequency components of noise that is not heart noise. Since these frequency components are similar, the heart noise sounding section detector 311 may detect both as heart noise sections. For this reason, the cardiac noise outline determination unit 312 detects cardiac noise based on the outline of the frequency components of the cardiac noise shown in FIG. 6 and the frequency components of the noise that is not the cardiac noise shown in FIG.

  The frequency component shown in FIG. 6 and FIG. 7 has a characteristic that the power decreases as the frequency increases. For this reason, the boundary line used as the upper limit of a frequency component can be calculated | required by using a predetermined threshold value. The shape of this boundary line is defined as the approximate shape of the frequency component.

  The cardiac noise outline determining process by the cardiac noise outline determining unit 312 will be described with reference to FIG. First, the cardiac noise outline determining unit 312 calculates the ratio of frequency components for each frame of the cardiac noise sounding interval detected by the cardiac noise sounding interval detecting unit 311 (step S121). Next, a rough shape is detected from the ratio of the frequency components calculated in step S121 (step S122).

  In the process of step S122, for example, the frequency component of the heart noise shown in FIG. 6 becomes smaller as the frequency approaches 1000 Hz. Furthermore, the amount of change in the ratio of frequency components is small above the upper limit frequency of the frequency components of noise that is not cardiac noise. The heart noise outline determining unit 312 uses this characteristic, and determines that the location is a boundary line when a ratio of a frequency component smaller than a predetermined threshold value continues for a predetermined number of times for each frame of the heart noise sounding interval. Furthermore, the outline of the heart noise sounding interval is obtained based on the boundary line determined here. In this case, the predetermined number of times is, for example, five times, but is not limited.

  The cardiac noise outline detection process in step S120 is not limited to the above-described example. As another example, a method in which a square value of a frequency component is obtained for each frame of a cardiac noise sounding interval and a boundary line is used when the difference between the frequency components is equal to or larger than a predetermined threshold value can be applied.

  Next, based on the outline detected in step S122, the cardiac noise outline determining unit 312 determines whether or not the frequency component in the cardiac noise sounding section is cardiac noise (step S123).

  In step S123, it is determined whether or not the frequency component of the heart noise sounding section is heart noise by expressing the outline of the heart noise sounding section detected in step S122 with a cubic approximation curve. FIG. 9 shows an approximate curve in the case of cardiac noise, which is a convex curve. FIG. 10 shows an approximate curve when there is no cardiac noise, and is an uneven curve.

  As described above, when the difference value with the adjacent frame is calculated based on the outline expressed as the cubic approximate curve, the sign of the difference value changes in the approximate curve where the approximate curve is uneven. Changes at one place, and changes at two places in FIG. Thus, if there is only one change in the sign of the difference value, it is determined as heart noise, and if there are two or more locations, it is determined not as heart noise.

  The determination of whether or not the heart noise is in step S123 is not limited to the above-described one. For example, the outline of the heart noise sounding section detected in step S122 is normalized to a predetermined number of frames, and the normalized outline is obtained. It is also possible to make a determination based on the difference value of the shape.

  In the process of step S123 based on the normalized outline difference value, first, a difference value between adjacent frames of the outline of the cardiac noise sounding section detected in step S122 is calculated. In this case, the number of frames in the heart noise sounding interval differs depending on the location of the heart noise and is not constant. Therefore, normalization is performed to compare features of outlines.

  As a normalization method, there is a method of unifying with a predetermined number of frames from the frame which is the center of the outline. For example, since the sound generation time of heart noise is generally set to 300 ms to 500 ms, a method of consolidating the frames for the sound generation time can be considered. The result is determined using the generalized normalization information summarized in this way.

  Such a normalization method will be described with reference to FIG. For example, as shown in FIG. 11A, when the approximate number of frames of the heart noise sounding section is n frames, the difference value from adjacent frames is n−1 frames as shown in FIG. Here, in order to unify the predetermined number of frames u, n−1 is the number of frames divisible by u. For this purpose, as in (C), the average of n-1 adjacent frames in (B) is obtained, and this number of frames is added by the number of frames a to obtain the number of frames divisible by u. The normalization method is an example, and the present invention is not limited to this.

  FIG. 12 and FIG. 13 are graphs showing information obtained by normalizing the outline of the cardiac noise sounding interval. FIG. 12 shows information obtained by normalizing a heart noise sounding interval that is a heart noise, and FIG. 13 shows information obtained by normalizing a heart noise sounding interval that is not a heart noise. In FIG. 12, it is a concave curve, and in FIG. 13, it is not a concave shape. Using this difference in shape, it is determined whether or not the frequency component in the heart noise sounding section is heart noise.

  For example, when the difference between adjacent normalized information in FIG. 12 is calculated, subtracting the first value from the second value results in a negative value. Similarly, when calculating up to the fifth, it becomes minus, minus, plus, plus, and it can be determined that it is a concave curve. In this way, if the normalized information is concave, it can be determined that the frequency component of the heart noise sounding section is heart noise, otherwise it is not heart noise.

  Furthermore, a determination method using the difference in variation with respect to the normalized information is also possible. For the comparison of variations, a standard deviation or a coefficient of variation, which is a numerical value indicating the degree of dispersion of statistical values or established variables, is used. A determination method using the difference in variation will be described with reference to FIGS.

  FIG. 14 is a graph of information obtained by normalizing a cardiac noise sound generation interval, which is cardiac noise, with the difference value between adjacent frames as the horizontal axis and the frequency as the vertical axis. FIG. 15 is a graph of information obtained by normalizing a cardiac noise sounding section that is not cardiac noise, with the difference value between adjacent frames as the horizontal axis and the frequency as the vertical axis. FIG. 14 shows that the variation of the heart noise is gathered around the standard deviation value 50, and FIG. 15 shows that the variation of the heart noise is widened.

  FIG. 16 is a graph in which the standard deviation of the normalized information of the heart noise sounding interval that is the heart noise and the normalized information of the heart noise sounding interval that is not the heart noise is obtained. In FIG. 16, the standard deviation of a heart noise sounding interval that is a heart noise is about 20, and the standard deviation of a heart noise sounding interval that is not a heart noise is about 45. The variation of the normalized information is large, and for example, the threshold value th with the standard deviation 30 as a threshold value can be classified as a boundary. That is, if the standard deviation of the normalized information is a value lower than the threshold th, it can be determined that the noise is cardiac noise, and if it is a high value, it is not the cardiac noise.

  Returning to FIG. 8, when it is determined in step S123 that it is not cardiac noise (step S123: No), the data of the cardiac noise sounding section determined not to be cardiac noise is deleted (step S124). If it is determined that the heart noise is present (step S123: Yes), the heart noise outline determining unit 312 determines whether the determination of the heart noise is completed (step S125).

  In step S125, if it is determined that the determination of cardiac noise has not ended (step S125: No), that is, if there is a cardiac noise sounding section that is a determination target of whether or not it is cardiac noise, The processing from step S121 is executed for the heart noise sounding section. When it is determined that the determination of the cardiac noise has been completed (step S125: Yes), this process ends.

  As described above, since the heart noise outline determination unit 312 determines the heart noise based on the outline of the heart noise sounding interval, even if the input heart sound signal includes noise other than the heart noise, the heart noise is accurately detected. Can be detected. For this reason, it is possible to detect heart noise with high accuracy at medical sites and home medical care where appropriate judgment is required.

  Returning to FIG. 3, the display information generation unit 400 generates display information for displaying the heart noise data on the display unit 130 based on the data of the sound generation section of the heart noise determined to be the heart noise in step S120 ( Step S130). The display information generation process in step S130 is the same as the technique disclosed in the contents filed by the present applicant as Japanese Patent Application No. 2011-173291.

  As a specific example, as shown in FIG. 17, display information 501 indicating heart noise is generated in the heart noise sounding interval on the time axis, and the state of the heart sound is schematically illustrated. The display information of the heart noise is not limited to the information shown in 501, but may be changed depending on the size of the heart noise or the difference in the sound generation position.

  In addition, the generated display information is displayed on the display unit 130. An example of the display screen 800 in the display unit 130 is shown in FIG. A display screen 800 shown in FIG. 18 includes a GUI (Graphical User Interface) for performing operations, and these GUI operations are accepted by a keyboard, a mouse, a touch panel, or the like. Further, the display screen 800 shown in FIG. 18 is an example, and it is only necessary that various characteristic portions of heart sounds including heart noise can be appropriately identified.

  In FIG. 18, an open button 801 selects a heart sound signal input to the signal input unit 110 when operated. The heart sound signal selected here is, for example, a heart sound signal input in real time by a microphone or the like, or a recorded heart sound signal stored in the storage unit 140.

  The play button 802 plays the heart sound to be analyzed when operated. The heart sound to be analyzed to be reproduced means that if a heart sound is input to the heart sound input unit 110 in real time, output of the heart sound is started, and if it is a recorded heart sound, the recorded data to be reproduced Start playing. In addition, the heart sound reproduced by the operation of the play button 802 is not a heart sound including all the various characteristic parts or noise components input by a microphone or the like, but only the various characteristic parts detected by the heart sound detecting unit 300. It may be.

  When the stop button 803 is operated, the reproduction of the heart sound is stopped. The repeat button 804 is operated to repeatedly reproduce a predetermined repetition period.

  The heart sound display area 805 displays a heart sound signal to be reproduced along the time axis. In addition, a playback position display bar 806 and a playback section 807 are displayed on the heart sound display area. A reproduction position display bar 806 indicates the position currently being reproduced in the heart sound display area 805. The playback section 807 can designate a section to be played in the heart sound display area 805. The playback section 807 is also a repeated playback section by operating the repeat button 804.

  The enlargement / reduction button 808 is operated to change the scale of the cardiac noise display area 805. The scale operated by the enlargement / reduction button 808 can be in the time axis (horizontal axis) direction or the amplitude axis (vertical axis) direction of the cardiac noise display region 805.

  The information display area 809 displays detailed information on the characteristic parts of each heart sound displayed schematically. As specific examples, the names of characteristic parts constituting heart sounds such as I sound and II sound, and, when heart noise is included, are the heart noise and the name and shape of the heart noise.

  Through the above-described processing, the heart noise is detected with high accuracy by performing the outline determination of the heart noise even for the speech component in which it is not clear whether or not it is the heart noise. Further, by displaying the detected heart noise and the like as described above, the user can easily and appropriately confirm the state of the input heart sound.

  Next, the III sound analysis processing by the III sound analysis signal generation unit 230 and the III sound detection unit 430 will be described. FIG. 19 is a block diagram illustrating the configuration of the III sound detection unit 330. The III sound detection unit 330 is configured as a function executed by the arithmetic processing unit 120. The III sound detection section determination unit 331 and the III sound generation section A detection unit 332 is provided.

  Next, the III sound detection process for detecting the III sound will be described with reference to FIGS. With respect to the III sound detection process, the III sound analysis signal generation unit 230 generates an III sound analysis signal, and the III sound detection unit 330 detects the III sound based on the III sound analysis signal. When the processing of FIG. 20 is executed, the heart sound information processing apparatus 100 can be mounted or executed with the III sound analysis module 430. Further, when the detection result of the I sound II sound is used for detecting the III sound, the I sound II sound analysis module 420 can be further mounted or executed.

  In the flowchart of FIG. 20, step S200 corresponds to the analysis signal detection process, and steps S210 to S240 correspond to the heart sound detection process.

  First, the III sound analysis signal generation unit 230 generates a III sound analysis signal from the heart sound signal input from the signal input unit 110 (step S200). In the process in step S200, the heart sound signal input from the signal input unit 110 is first divided into frames having a predetermined time length. The predetermined time length in this case is 12 ms as an example, but is not limited to this value as long as the peak positions of the II sound and the III sound can be distinguished.

  Next, the heart sound signal divided in a frame of a predetermined time length is converted into a frequency domain using FFT or the like, and the power for each frequency component is calculated in each frame. In the present embodiment, the sum of the power of each frequency component of 0 to 300 Hz is mainly extracted and used as the III sound analysis signal.

  Next, the III sound detection unit 330 detects a time interval based on the sound generation position of the I sound and the II sound from the detection information of the I sound and the II sound detected by the I sound II sound detection unit 320 (step S210).

  Here, the time intervals of the I, II, and III sounds will be described with reference to FIGS. FIG. 21 shows an example of a heart sound in congestive heart failure, and FIG. 22 shows an example of a heart sound in acute myocardial infarction. FIG. 23 is a graph showing one cycle of heart sounds in congestive heart failure, and FIG. 24 is a graph showing one cycle of heart sounds in acute myocardial infarction. The heart sounds of congestive heart failure shown in FIG. 21 and FIG. 23 have clear peaks of I sound, II sound and III sound, but the heart sounds of acute myocardial infarction shown in FIG. 22 and FIG. Since the time interval is short, the II sound and the III sound cannot be clearly separated, and the III sound may be erroneously acquired as the II sound.

  Next, the III sound detection unit 330 performs clustering on the time interval between the I sound and the II sound detected in Step S210 (Step S220). As a clustering method, various methods such as a known k-means method can be applied.

  Next, the III sound detection section determination unit 331 determines the III sound detection section based on the clustering result in step S220 (step S230). The clustering in step S220 can be visualized as shown in FIGS. 25 and 26, for example, using a histogram of detected time intervals. FIG. 25 is an example of the clustering result based on the heart sound of congestive heart failure, and FIG. 26 is an example of the clustering result based on the heart sound of acute myocardial infarction.

  In FIG. 25 and FIG. 26, the left side of the threshold value th indicated by the broken line is the cluster 1 that is a cluster with a short time interval, and the right side is the cluster 2 that is a cluster with a long time interval. In the clustering result based on the heart sound of congestive heart failure shown in FIG. 25, the peaks of the I sound and the II sound are clear, and the frequency of each cluster is almost equal and takes the mode value. For this reason, it can be determined that the sound after the II sound is the III sound detection section. In the clustering result based on the heart sound of acute myocardial infarction shown in FIG. 26, there is a possibility that the sound where the time interval between the sound I and sound II is not the mode value is determined as sound II. For this reason, the part which has deviated from the mode value is set to the III sound detection section before the II sound.

  Returning to FIG. 20, based on the result of the III sound detection section determined in step S230, the III sound generation section detection unit 332 detects the III sound generation section (step S240). Generally, it is said that the III sound can be heard around 0.15 seconds after the II sound. For this reason, in this embodiment, based on the result of the III sound detection section determination process in step S230, the section between 0.13 seconds and 0.17 seconds before and after the II sound is set as the detection target range of the III sound generation section.

  In the process of step S240, the III sound is detected in the interval from 0.13 seconds to 0.17 seconds before and after the II sound that is the detection target range using the III sound analysis signal. Here, the reason for using the III sound analysis signal will be described with reference to FIG. FIG. 27 is a line graph showing the acute myocardial infarction signal of FIG. 22 as a III sound analysis signal. In FIG. 24, the acute myocardial infarction signal is represented by an analysis signal created with a frame length longer than the III sound analysis signal. In FIG. 24, the interval between the II sound and the III sound is short. Although it was not known whether it was a III sound, since the III sound analysis signal is a frame having a fine time length, the peaks of the II sound and the III sound can be discriminated as shown in FIG.

  In this way, by using the III sound analysis signal calculated in a frame with a fine time length, it becomes possible to detect the peak position even if the positions of the II sound and the III sound are close, so the III sound analysis signal Is used. When a peak is detected before the II sound using the III sound analysis signal, it is determined that the detected peak is the II sound generation section, and the II sound detected by the I sound II sound detection unit 320 is detected. Is corrected to III sound. If a peak is detected after the II sound, it is determined that the detected peak is a III sound.

  Next, the display information generation unit 400 generates display information for displaying the III sound data on the display unit 130 (step S250). The processing and display screen of step S250 are the same as the contents described in step S130 of FIG. FIG. 28 is a schematic diagram showing the III sound. In FIG. 28, the III sound 601 is shown after the I sound and the II sound. In this way, the position of the III sound can be grasped by displaying the I sound, the II sound, and the III sound.

  As described above, even when the peak of the III sound is not clear, the III sound is detected with high accuracy, and the user can appropriately grasp the state of the III sound. Further, these methods are not limited to the detection of the III sound but can be applied to the detection of other characteristic sounds constituting the heart sound.

Claims (5)

A heart noise analysis signal generator for generating a heart noise analysis signal for analyzing heart noise constituting the heart sound;
A cardiac noise detection unit for detecting cardiac noise based on the cardiac noise analysis signal generated by the cardiac noise analysis signal generation unit;
The cardiac noise detection unit includes a cardiac noise sounding interval detection unit that detects a cardiac noise sounding interval using the cardiac noise analysis signal, and the cardiac noise analysis according to a rough shape indicating a temporal change of the cardiac noise in the cardiac noise sounding interval. A cardiac noise outline determining unit that determines whether the signal is cardiac noise;
A heart sound information processing apparatus characterized by that. A heart noise analysis signal generator for generating a heart noise analysis signal for analyzing heart noise constituting the heart sound ;
A cardiac noise detection unit for detecting cardiac noise based on the cardiac noise analysis signal generated by the cardiac noise analysis signal generation unit;
The cardiac noise detector includes a cardiac noise sounding interval detector that detects a cardiac noise sounding interval using the cardiac noise analysis signal, and the cardiac noise analysis signal is converted into a cardiac noise according to a rough shape of the cardiac noise in the cardiac noise sounding interval. Comprising a cardiac noise outline determining unit for determining whether or not
The cardiac noise outline determination unit determines whether the cardiac noise analysis signal is cardiac noise based on the outline of a boundary between a frequency component of cardiac noise and a frequency component in which no cardiac noise exists in the cardiac noise sounding section. heart sound information processing apparatus wherein the determining. A heart noise analysis signal generator for generating a heart noise analysis signal for analyzing heart noise constituting the heart sound ;
A cardiac noise detection unit for detecting cardiac noise based on the cardiac noise analysis signal generated by the cardiac noise analysis signal generation unit;
The cardiac noise detector includes a cardiac noise sounding interval detector that detects a cardiac noise sounding interval using the cardiac noise analysis signal, and the cardiac noise analysis signal is converted into a cardiac noise according to a rough shape of the cardiac noise in the cardiac noise sounding interval. Comprising a cardiac noise outline determining unit for determining whether or not
The cardiac noise outline determination unit is based on a variation in normalized information at a boundary between a frequency component of cardiac noise and a frequency component in which no cardiac noise exists in the cardiac noise sounding section, and the cardiac noise analysis signal is cardiac noise heart sound information processing apparatus you wherein the determining whether. On your computer,
A heart noise analysis signal generation step for generating a heart noise analysis signal for analyzing the heart noise constituting the heart sound;
A cardiac noise sounding interval detection step for detecting a cardiac noise sounding interval using the cardiac noise analysis signal generated in the cardiac noise analysis signal generation step;
Whether or not the cardiac noise analysis signal is cardiac noise based on the outline of the boundary between the frequency component of cardiac noise and the frequency component in which cardiac noise does not exist in the cardiac noise speech zone detected in the cardiac noise speech zone detection step A cardiac noise outline judging step for judging whether
A program characterized by having executed. On your computer,
A heart noise analysis signal generation step for generating a heart noise analysis signal for analyzing the heart noise constituting the heart sound;
A cardiac noise sounding interval detection step for detecting a cardiac noise sounding interval using the cardiac noise analysis signal generated in the cardiac noise analysis signal generation step;
Based on the variation of the normalized information at the boundary between the frequency component of the cardiac noise and the frequency component in which no cardiac noise exists in the cardiac noise generation interval detected in the cardiac noise generation interval detection step, the cardiac noise analysis signal is a cardiac noise. A cardiac noise outline determining step for determining whether or not there is,
A program characterized by having executed.

Images