JP5432848B2 - Pulse photometer and pulse rate identification method - Google Patents

Description

  The present invention relates to a signal processing method for processing two signals to extract a common signal component, a pulse rate specifying method in a pulse photometer, and a pulse photometer using the same.

  An oximeter is known as a pulse photometer that can measure the oxygen saturation of arterial blood. The oximeter includes a probe having a light emitting diode that emits light having different wavelengths of infrared light and red light, and a photodiode that outputs an electric signal in accordance with the received light. The probe is attached to the subject's fingertip or earlobe, light of a different wavelength is intermittently emitted from the light emitting diode to the living tissue of the subject, and transmitted light or reflected light from the living tissue is received by the photodiode to generate an electrical signal ( (Pulse wave) is obtained and the absorbance of the living tissue is measured.

  Oxyhemoglobin contained in a large amount in arterial blood absorbs infrared light (wavelength 940 nm) well, and reduced hemoglobin contained in a large amount in venous blood absorbs red light (wavelength 660 nm) well. Therefore, the relative concentrations of oxyhemoglobin and deoxyhemoglobin can be measured by measuring light having these wavelengths transmitted or reflected from the living tissue. In addition, arterial blood containing a large amount of oxyhemoglobin is pumped out by the heartbeat (beat). Therefore, when the absorbance is measured, the pulse rate can be measured from the change in absorbance.

  In such an oximeter, when a body movement such as movement of the subject occurs during measurement, a frequency other than the pulse is mixed in the electrical signal. The amplitude of the electrical signal due to this body movement is often the same as or larger than the amplitude due to the pulse, and the frequency with the maximum amplitude in the frequency spectrum of the electrical signal can simply be identified as the pulse rate. Therefore, it was difficult to specify the pulse rate.

  Therefore, as described in Patent Document 1 and the like, various attempts have been made to reduce the noise included in the electrical signal to facilitate the specification of the pulse rate and to increase the reliability of the specified pulse rate. It has been made conventionally.

  By the way, when the pulse rate is specified from the electric signal with reduced noise as described above, the pulse rate is specified as follows. First, assuming that the pulse rate does not fluctuate greatly instantaneously, it is assumed that the pulse rate at time t is within ± 10% of the pulse rate at time t−1 immediately before it, and is obtained at time t. Of the frequency spectrum of the obtained electrical signal, only the frequency spectrum within the range of ± 10% of the pulse rate at time t-1 was considered, and the frequency with the maximum spectrum peak was identified as the pulse at time t.

Japanese Patent No. 4352315 Japanese Patent No. 4196209

  However, in the above-described pulse rate specifying method, when the pulse rate fluctuates by 10% or more from the immediately preceding pulse rate due to arrhythmia or the like, the frequency to be a pulse is overlooked, or artifact (noise) is superimposed on the electrical signal. However, when the pseudo spectrum peak has a larger amplitude than the true spectrum peak, there is a possibility that an erroneous frequency is specified as the pulse rate.

  Accordingly, an object of the present invention is to provide a highly reliable pulse rate specifying method and a pulse photometer using the same without overlooking the frequency to be set as a pulse.

In order to achieve the above object, the present invention provides the following.
(1) a light emitting unit that irradiates a living tissue with light of two different wavelengths;
In a pulse photometer comprising a light receiving unit that generates a signal in response to light generated from the light emitting unit and transmitted or reflected through a living tissue,
A frequency spectrum conversion unit for obtaining frequencies and amplitudes at all peaks of the frequency spectrum from a frequency spectrum obtained by frequency analysis of the signal measured in a predetermined time zone;
A frequency correlation degree calculating unit for calculating a frequency correlation degree for each peak from a difference between the frequency at all the peaks of the frequency spectrum and a previous pulse frequency which is a pulse frequency immediately before the predetermined time zone;
An amplitude correlation calculating unit that calculates an amplitude correlation for each peak from the amplitude at all peaks of the frequency spectrum;
A total correlation calculating unit for calculating a total correlation for each peak from the frequency correlation for each peak and the amplitude correlation for each peak;
A pulse photometer comprising a pulse rate specifying unit that specifies a pulse rate in the predetermined time zone based on the total correlation.
(2) The pulse photometer according to (1), wherein the pulse rate specifying unit specifies the frequency of the peak at which the total degree of correlation is maximum as the pulse rate in the predetermined time zone.
(3) The pulse photometer according to any one of (1) and (2), wherein the pulse photometer uses the specified pulse frequency as the next previous pulse frequency.
(4) The frequency correlation calculation unit includes:
Obtain the absolute value of the difference between the frequency and the previous pulse frequency at all peaks of the frequency spectrum,
Normalizing the absolute value of the difference so that the maximum absolute value of the difference is 1,
The pulse photometer according to any one of (1) to (3), wherein a value obtained by subtracting the normalized absolute value of the difference from 1 is used as the frequency correlation.
(5) The amplitude correlation degree calculation unit sets a value obtained by multiplying a value obtained by normalizing the amplitude so that a maximum value of the amplitude is 1 and a weighting coefficient as the amplitude correlation degree. The pulse photometer according to any one of 1) to (4).
(6) The pulse photometer according to any one of (1) to (5), wherein the total correlation degree is a sum of a square of the frequency correlation degree and a square of the amplitude correlation degree.
(7) The pulse photometer
The pulse photometer according to any one of (1) to (6), further comprising a preprocessing unit that reduces noise from the two signals from the light receiving unit corresponding to light of two different wavelengths.
(8) The pulse photometer according to (7), wherein the preprocessing unit applies at least one of a rotation method and a double rotation method to the two signals from the light receiving unit.

According to the present invention, the following is provided to achieve the above object.
(9) a light emitting unit that irradiates a living tissue with light of two different wavelengths;
In a pulse photometer comprising a light receiving unit that generates a signal in response to light generated from the light emitting unit and transmitted or reflected through a living tissue,
A frequency spectrum conversion step of obtaining frequencies and amplitudes at all peaks of the frequency spectrum from a frequency spectrum obtained by frequency analysis of the signal measured in a predetermined time zone;
A frequency correlation calculating step of calculating a frequency correlation for each peak from a difference between the frequency at all the peaks of the frequency spectrum and a previous pulse frequency that is a pulse frequency immediately before the predetermined time zone;
An amplitude correlation calculating step of calculating an amplitude correlation for each peak from the amplitude at all the peaks of the frequency spectrum;
A comprehensive correlation calculating step of calculating a total correlation for each peak from the frequency correlation for each peak and the amplitude correlation for each peak;
A pulse rate specifying method comprising a pulse rate specifying step of specifying a pulse rate in the predetermined time zone based on the total correlation degree, wherein the steps are executed in order.
(10) The pulse rate specifying method according to (9), wherein, in the pulse rate specifying step, the frequency of the peak at which the total correlation degree is maximum is specified as the pulse rate in the predetermined time zone.
(11) The pulse rate specifying method according to any one of (9) and (10), wherein the pulse photometer sets the frequency of the specified pulse rate as the next previous pulse frequency.
(12) In the frequency correlation calculation step,
Obtain the absolute value of the difference between the frequency and the previous pulse frequency at all peaks of the frequency spectrum,
Normalizing the absolute value of the difference so that the maximum absolute value of the difference is 1,
The pulse rate specifying method according to any one of (9) to (11), wherein a value obtained by subtracting the normalized absolute value of the difference from 1 is used as the frequency correlation degree.
(13) In the amplitude correlation calculation step, a value obtained by multiplying a value obtained by normalizing the amplitude so that a maximum value of the amplitude is 1 and a weighting coefficient is used as the amplitude correlation. The pulse rate specifying method according to any one of 9) to (12).
(14) The pulse rate identification method according to any one of (9) to (13), wherein the total correlation is the sum of the square of the frequency correlation and the square of the amplitude correlation.
(15) A preprocessing step of reducing noise from the two signals from the light receiving unit corresponding to light of two different wavelengths is executed before the frequency spectrum conversion step (9) to (9) The pulse rate specifying method according to any one of (14).
(16) The pulse rate identification method according to (15), wherein, in the preprocessing step, at least one of a rotation method and a double rotation method is applied to the two signals from the light receiving unit.

  According to the pulse photometer and the pulse rate specifying method according to the present invention described above, all the peaks of the frequency spectrum obtained by frequency analysis of the signal obtained in a predetermined time zone are considered. Therefore, since the frequency that should be a pulse is not overlooked, the pulse rate can be specified with high reliability. In addition, since the pulse rate is specified not only by the frequency but also by the total degree of correlation taking into account the amplitude indicating the signal strength, the reliability of the obtained pulse rate is high.

It is a schematic block diagram of the oximeter which can apply the pulse rate specific method which concerns on embodiment of this invention. It is a functional block diagram of the process part in the oximeter of FIG. It is a flowchart of the process performed by the process part of FIG. It is a time chart of signal IR by infrared light and signal IR by red light R. It is a figure which shows the spectrum which FFT-analyzed the signal shown in FIG. It is a schematic diagram showing the concept of total correlation degree Dis. It is a figure which shows the time passage of the pulse rate obtained by applying the pulse rate measuring method by embodiment of this invention to signal IR. It is a figure which shows the time passage of the pulse rate obtained by applying the conventional pulse rate measuring method to signal IR. It is a figure which shows the time passage of the pulse rate obtained by applying the pulse rate measuring method by embodiment of this invention to the signal Z1 which performed the pre-processing. It is a figure which shows the time passage of the pulse rate obtained by applying the conventional pulse rate measuring method to the signal Z1 which performed the pre-processing.

  Hereinafter, a pulse photometer and a pulse rate specifying method according to an embodiment of the present invention will be described using an oximeter, which is a kind of pulse photometer, as an example.

<Overall structure of oximeter>
An oximeter measures arterial oxygen saturation and pulse rate by irradiating a living tissue with light of two types of wavelengths having different absorbances and measuring the transmitted light or reflected light. The pulse rate specifying method according to this embodiment specifies the pulse rate from the signal acquired by the light receiving element of the oximeter and displays it on the display unit.

  FIG. 1 shows a schematic configuration diagram of an oximeter. The light emitted alternately and continuously from the infrared light emitting element 1 and the red light emitting element 2 (light emitting unit) driven by the drive circuit 3 passes through the living tissue 4 and the transmitted light is transmitted through the photodiode ( The light is received by a light receiving unit 5. Note that light having a wavelength of 940 nm is emitted from the infrared light emitting element 1 as infrared light, and light having a wavelength of 660 nm is emitted from the red light emitting element 2 as red light. In addition, you may comprise so that the photodiode 5 may receive reflected light instead of transmitted light.

  The photodiode 5 outputs an electrical signal corresponding to the amount of received light and is amplified by the amplifier 6. The electric signal amplified by the amplifier 6 is separated into a signal by infrared light and a signal by red light by a multiplexer 7 and is filtered by filters 8-1 and 8-1 corresponding to the respective signals to reduce noise. The signal inputted to the A / D converter 9 is digitized, and a signal IR by infrared light and a signal R by red light are obtained.

  Each of the digitized signals IR and R is input to the processing unit 10, and the processing unit 10 is converted into a frequency spectrum conversion unit 10a, a frequency correlation calculation unit 10b, an amplitude correlation as shown in FIG. The pulse rate specified by the processing unit 10 is displayed on the display unit 11 by functioning as the degree calculation unit 10c, the comprehensive correlation calculation unit 10d, and the pulse rate specifying unit 10e and performing the signal processing shown in FIG.

<Identification of pulse rate>
Hereinafter, the pulse rate specifying method executed by the processing unit 10 will be described in detail. First, the processing unit 10 receives the infrared light and red light transmitted from the living tissue 4 as described above by the light receiving element 6 and acquires the signals IR and R output as electric signals (step S0).

  FIG. 4 shows a time chart of the signal IR and the signal R acquired by the processing unit 10. The A / D converter 9 samples these signals IR and R at a sampling period of 16 ms, and FIG. 4 shows a set of 1000 signals. In FIG. 4, a solid line represents an infrared light signal IR, a two-dot chain line represents a red light signal R, the horizontal axis represents time [s], and the vertical axis represents amplitude (signal intensity) [mV].

  The signal IR obtained by the A / D converter 9 is input to the processing unit 10. Based on the signal IR, the processing unit 10 calculates the total correlation degree Dis from the frequency correlation degree ld and the amplitude correlation degree a and specifies the pulse rate as described in detail below. In the following description, the word “pulse frequency [Hz]” is used to simplify the description. The relationship of pulse rate [times / minute] = pulse frequency [Hz] × 60 is established. Hereinafter, the pulse rate specifying method according to the present embodiment will be described in detail with reference to FIGS.

The pulse rate specifying method described below is a method found by the following findings (1) to (3) by the inventors.
(1) The pulse frequency M (t) at time t is one of a plurality of peaks appearing in the FFT spectrum of the signal on page Pt at time t.
(2) The pulse frequency M (t) at time t is likely to be a value in the vicinity of the pulse frequency M (tpr) at time tpr immediately before that.
(3) Even if the frequency is far from the pulse frequency M (tpr) at the time tpr immediately before that, if the amplitude is large, the possibility of the pulse frequency M (t) at the time t is high.

  Note that the page P here is, for example, a collection of 1000 points sampled at a sampling period of 16 ms, that is, a collection of sample points for 16 [s]. For example, the time chart of FIG. 4 shows a signal for 16 seconds as a signal for one page. The page Pt at the time t is a set of 1000 sampling points of the signal between the time t-16 [s] and the time t [s]. Note that the sampling period is not limited to 16 ms, and may be set to an arbitrary sampling period.

  First, the signal IR output from the A / D converter 9 is input to the frequency spectrum conversion unit 10a. The frequency spectrum conversion unit 10a obtains the FFT spectrum of the signal IR on the page Pt at time t (step S1), and specifies the peak frequency L and amplitude A appearing in the FFT spectrum (step S2). Specifically, the frequency spectrum conversion unit 10a obtains the FFT spectrum of the signal IR on the page Pt at time t as shown in FIG. 5, identifies all the peaks included in this spectrum, and the frequency of each peak. (L1 to Ln) and amplitudes (A1 to An) are acquired.

  Based on knowledge (1), any one of these identified peak frequencies L should be the pulse frequency M (t) at time t. Note that although all peaks are evaluated, it is not necessary to consider a frequency that cannot be a pulse frequency. In consideration of the pulse rate of the human body, for example, it is sufficient to consider a frequency of 0.5 to 7 [Hz] (30 to 420 times / minute in terms of pulse rate) as shown in FIG.

  Next, the peak frequency L of the output from the frequency spectrum conversion unit 10a is input to the frequency correlation degree calculation unit 10b. Based on the knowledge (2), the frequency correlation calculation unit 10b calculates the pulse frequency Lo = M (tpr) at time tpr immediately before time t for all the peak frequencies L included in the page Pt at time t (hereinafter referred to as “t”). The distance from the previous pulse frequency Lo) is evaluated (step S3). In the present embodiment, the time tpr immediately before the time t is set as a time before the time t [s] by 16 ms which is a sampling period.

The frequency correlation calculation unit 10b obtains an absolute value La of the difference between the frequency L = (L1, L2, L3,..., Ln) of each peak acquired by the frequency spectrum conversion unit 10a and the previous pulse frequency Lo.
La = | L-Lo | = (| L1-Lo |, | L2-Lo |, | L3-Lo |,..., | Ln-Lo |)

  According to knowledge (2), the frequency closer to the previous pulse frequency Lo is more likely to be the pulse frequency M (t) at time t. However, since the index La obtained here shows a larger value as the frequency is farther from the previous pulse frequency Lo, the frequency correlation degree calculation unit 10b shows the frequency correlation degree so that the closer to the previous pulse frequency Lo, the larger the value is. Id is calculated as follows.

First, each La is divided by the maximum value of La so that the largest difference among La is 1, normalization la is obtained, and a value obtained by subtracting the obtained la from 1 is defined as a frequency correlation degree ld.
la = La / max (La)
ld = 1-la

  The frequency correlation degree ld of each peak output from the frequency correlation degree calculation unit 10b obtained in this way takes a value in the range of 0 to 1, and the frequency having the frequency correlation degree ld close to 1 is the previous one. It is close to the pulse frequency Lo, indicating that there is a high possibility that it is the pulse frequency M (t) at time t. However, when the pulse suddenly changes depending on the condition of the subject (for example, arrhythmia), the pulse rate at time t is not always close to the previous pulse frequency.

  Therefore, when the pulse suddenly changes, attention is paid to the fact that the amplitude shows a large value even if the frequency changes suddenly. Based on the knowledge (3), the frequency having a large amplitude is also set to the pulse frequency M ( The amplitude correlation degree a is obtained by the amplitude correlation degree calculation unit 10c so as to be considered as a candidate for t).

The amplitude correlation calculation unit 10c receives the amplitude A from the output of the frequency spectrum conversion unit 10a, and evaluates the signal strength of all peaks included in the page Pt at time t. Specifically, the amplitude correlation calculation unit 10c performs frequency spectrum conversion on the amplitudes A = (A1, A2, A3,..., An) of all peaks appearing in the FFT spectrum of the signal Z1 on the page Pt at time t. Obtained from the unit 10a. Next, the amplitude correlation calculation unit 10c is normalized by dividing each A by the maximum value max (A) of the amplitude A so that the maximum value of the amplitudes A becomes 1, as shown in the following formula. The value A / max (A) is multiplied by the weighting coefficient w to output a normalized amplitude a = (a1, a2, a3,..., An) (step S4).
a = w · (A / max (A))

  The weighting coefficient w in the above formula determines how much the amplitude correlation degree a is reflected in the total correlation degree Dis with respect to the frequency correlation degree ld when the total correlation degree Dis is obtained. That is, the balance between knowledge (2) and (3) is determined by the weighting coefficient w. In this embodiment, 0.7 is set to w.

  Next, in consideration of the findings (2) and (3), the total correlation calculation unit 10d may have the pulse frequency M (t) at the time t for all frequencies included in the page Pt at the time t. The total correlation degree Dis representing the height is acquired.

  In this embodiment, as shown in FIG. 6, when each peak is plotted on a graph in which the horizontal axis is the frequency correlation degree ld and the vertical axis is the amplitude correlation degree a, the total distance as the distance between the plot point and the origin O is obtained. It is considered that the degree of correlation Dis appears. The greater the distance, the greater the overall correlation degree Dis and the higher the possibility that the pulse frequency M (t) at time t.

Therefore, the frequency correlation degree ld from the frequency correlation degree calculation section 10b and the amplitude correlation degree a from the amplitude correlation degree calculation section 10c are input to the total correlation degree calculation section 10d, and the total correlation degree Dis is set as a frequency as follows: The square root of the sum of the square of the correlation degree ld and the square of the amplitude correlation degree a is output (step S5).
Dis = {ld ² + a ² } ^1/2

The total correlation degree Dis of each peak obtained by the total correlation degree calculation unit 10d is input to the pulse rate specifying unit 10e, and the pulse rate specifying unit 10e selects a frequency indicating the largest Dis from the total correlation degree Dis at the time t. The pulse frequency is identified and the pulse frequency M (t) is output (step S6).
M (t) = max (Dis)

  In the above embodiment, the weighting coefficient w is set to about 0.7, but the weighting coefficient w can be set to an arbitrary value. If w = 1 is set, the frequency and the amplitude have the same influence on the overall correlation, and the frequency having the same amplitude as the frequency at the immediately preceding time is set to the pulse frequency M (t ) Will be specified. Setting w in this way is effective when it is assumed in advance that the pulse of the subject greatly fluctuates. Conversely, if w is set to a value close to 0, the degree to which the pulse frequency M (t) at time t is determined by the frequency rather than the amplitude can be increased.

  Although the weighting coefficient w can be set to a value larger than 1, if set in this way, the amplitude is more important than the frequency, and noise such as body movement having a larger amplitude than the pulse is likely to occur. It is not practical to identify the pulse on the oximeter. Therefore, normally, the weighting coefficient w is set from the range of 0 to 1. Also, the weighting coefficient w can be a fixed value such as 0.7 as in the present embodiment, or w can be changed linearly or curvedly according to the frequency.

  The weighting coefficient w described above may be set in advance in the program, or an operation unit for the weighting coefficient w may be provided in the oximeter so that the weighting coefficient w can be set for each measurement.

  Of the frequency peaks obtained from the signal Z1 of the page Pt at time t through the above steps S0 to S6, the frequency peak in the vicinity of the previous pulse frequency Lo and having a large amplitude is the pulse frequency M ( t). The pulse frequency M (t) specified in this way is multiplied by 60 and inputted to the display unit 11 from the pulse rate specifying unit 10e of the processing unit 10, and the pulse frequency M (t) multiplied by 60 is displayed as the pulse rate. It is displayed on the part 11 (step S7).

  For the total correlation degree Dis obtained as described above, each total correlation degree Dis is normalized by the total sum of the total correlation degrees Dis for each frequency to obtain nor (Dis) as a probability density function, and this nor (Dis) The frequency indicating the maximum value may be specified as the pulse frequency M (t) of the current page Pt.

  In the above example, the overall correlation degree Dis is defined as the square root of the sum of the square of the frequency correlation degree ld and the square of the amplitude correlation degree a, but the sum of the square of the frequency correlation degree ld and the square of the amplitude correlation degree a. Alternatively, the total correlation degree may be evaluated as the sum of the frequency correlation degree ld and the amplitude correlation degree a.

  After the pulse rate is specified by the above steps S0 to S7, the pulse rate specified this time is used as the previous pulse frequency Lo for the next process of specifying the pulse rate (for example, time t + 16 [ms]) as the previous pulse frequency Lo. The total correlation degree Dis is obtained and the pulse rate can be identified with high reliability continuously.

  The pulse frequency immediately after the start of measurement may be specified using any known method. For example, (1) The pulse frequency is defined as the peak frequency from within the range of frequencies that can be taken as fluctuations in the pulse rate. (2) If the pulse frequency is statistically the basic frequency of the pulse rate, it appears as a spectrum value. There is a method in which a sufficient threshold value is set and the pulse rate is obtained from the lowest frequency among the frequencies exceeding the threshold value.

<Effect of pulse rate identification method using comprehensive correlation>
7 and 8, the effect of the pulse rate specifying method according to the present embodiment will be confirmed by comparing with the conventional method.

  FIG. 7 shows the time lapse of the pulse rate specified using the total correlation degree Dis as described above. On the other hand, FIG. 8 shows a frequency spectrum having a maximum peak of the frequency spectrum from the frequency spectrum in the range of ± 10% of the frequency corresponding to the pulse rate at time tpr immediately before time t, as in the prior art. It shows the time transition of the pulse rate when specified as the pulse rate in.

  The rapid fluctuation of the pulse rate shown in FIG. 8 is caused by the fact that a large amplitude frequency due to body movement or the like is mixed as noise within a range of ± 10% of the frequency corresponding to the previous pulse rate. It seems that the frequency of large amplitude has been identified as the pulse frequency.

  However, as in the pulse rate identification method according to the present embodiment shown in FIG. 7, if both the degree of closeness of frequency and the degree of amplitude are taken into account, rapid fluctuations in pulse rate are extremely small, and body motion It can be confirmed that the pulse rate can be specified with high reliability without being confused by noise of large amplitude.

  Further, according to the pulse rate specifying method according to the present embodiment, not considering only the frequency within ± 10% from the previous pulse frequency, but considering all frequencies as pulse frequency candidates, the original pulse frequency is There is little risk of oversight, and a highly reliable pulse rate can be displayed. Moreover, as described above, the pulse rate specifying method according to the present embodiment can specify the pulse mainly by addition and subtraction, and does not require complicated calculation, so the burden of calculation processing does not increase and the oximeter is inexpensive. Can be configured.

  As described above, the oximeter using the pulse rate specifying method according to the present embodiment can display a more reliable pulse rate.

  In addition, according to the pulse rate specifying method according to the present embodiment, since the frequency correlation degree a greatly affects the overall correlation degree Dis, the specified pulse rate is greatly influenced by the pulse rate immediately before the specification. Therefore, immediately after the start of measurement (the first and second plots in FIG. 7), the pulse rate may not always be measured accurately if appropriate settings are not made because the pulse rate greatly depends on the initial value. .

  Therefore, if the noise component is previously removed from the signal input to the frequency spectrum conversion unit 10a of the processing unit 10, a highly reliable pulse rate can be continuously specified from the initial stage. Specifically, the signal from the A / D converter 9 may be output to the preprocessing unit, and the signal whose noise is reduced by the preprocessing unit may be input to the frequency spectrum conversion unit 10a. The pre-processing unit can execute a known noise reduction method known to those skilled in the art including a rotation method (Patent Document 1) and a double rotation method (Patent Document 2).

  FIG. 9 shows an example of applying the rotation method described in Patent Document 1 to the signal IR and the signal R as preprocessing, applying the pulse rate specifying method according to the present embodiment to the obtained signal Z1, and obtaining the obtained pulse rate. It is the figure which showed time passage.

  FIG. 10 shows, as a comparative example, the rotation method described in Patent Document 1 is applied to the signal IR and the signal R as preprocessing, and the obtained signal Z1 is ± 10% of the signal IR at the time tpr immediately before the conventional time t. The time transition of the pulse rate when the frequency having the maximum peak in the frequency spectrum is identified as the pulse rate at time t is shown.

  First, FIG. 7 in which the pulse rate specifying method according to this embodiment is applied to a signal IR that has not been subjected to preprocessing, and FIG. 8 in which the pulse rate specifying method according to this embodiment is applied to a signal Z1 that has been subjected to preprocessing. Are compared, it is confirmed that the rapid fluctuation depending on the initial value of the pulse rate immediately after the start of the measurement is suppressed, and a highly reliable pulse rate can be specified from the initial stage.

  Further, in FIG. 10 in which the conventional method for specifying the pulse rate is applied to the signal Z1 obtained by simply using the rotation method as the preprocessing, a rapid fluctuation of the pulse rate is seen, but the preprocessed signal Z1 is applied to the signal Z1. In FIG. 9 to which the pulse rate specifying method according to this embodiment is applied, it is confirmed that the pulse rate is not rapidly changed and a highly reliable pulse rate can be specified continuously.

1, 2 :: Light emitting element, 4: Biological tissue, 5: Light receiving element, 6: Amplifier, 7: Multiplexer, 8-1, 8-2: Filter, 9: A / D converter, 10: Processing unit, 11 : Display unit, 12: ROM, 13: RAM, IR: Signal by infrared light, R: Signal by red light, ld: Frequency correlation, a: Amplitude correlation, Dis: Total correlation

Claims (16)

A light emitting unit that irradiates a living tissue with light of two different wavelengths;
In a pulse photometer comprising a light receiving unit that generates a signal in response to light generated from the light emitting unit and transmitted or reflected through a living tissue,
A frequency spectrum conversion unit for obtaining frequencies and amplitudes at all peaks of the frequency spectrum from a frequency spectrum obtained by frequency analysis of the signal measured in a predetermined time zone;
A frequency correlation degree calculating unit for calculating a frequency correlation degree for each peak from a difference between the frequency at all the peaks of the frequency spectrum and a previous pulse frequency which is a pulse frequency immediately before the predetermined time zone;
An amplitude correlation calculating unit that calculates an amplitude correlation for each peak from the amplitude at all peaks of the frequency spectrum;
A total correlation calculating unit for calculating a total correlation for each peak from the frequency correlation for each peak and the amplitude correlation for each peak;
A pulse photometer comprising a pulse rate specifying unit that specifies a pulse rate in the predetermined time zone based on the total correlation.   2. The pulse photometer according to claim 1, wherein the pulse rate specifying unit specifies the frequency of the peak at which the total degree of correlation is maximized as the pulse rate in the predetermined time zone.   The pulse photometer according to claim 1, wherein the pulse photometer sets the frequency of the specified pulse rate as the next previous pulse frequency. The frequency correlation calculation unit includes:
Obtain the absolute value of the difference between the frequency and the previous pulse frequency at all peaks of the frequency spectrum,
Normalizing the absolute value of the difference so that the maximum absolute value of the difference is 1,
4. The pulse photometer according to claim 1, wherein a value obtained by subtracting the normalized absolute value of the difference from 1 is used as the frequency correlation degree.   The amplitude correlation degree calculating unit sets the amplitude correlation degree to a value obtained by multiplying a value obtained by normalizing the amplitude so that a maximum value of the amplitude becomes 1 and a weighting coefficient. 5. The pulse photometer according to any one of 4 above.   6. The pulse photometer according to claim 1, wherein the total correlation degree is a sum of a square of the frequency correlation degree and a square of the amplitude correlation degree. The pulse photometer is
The pulse photometer according to any one of claims 1 to 6, further comprising a preprocessing unit that reduces noise from the two signals from the light receiving unit corresponding to light having two different wavelengths.   8. The pulse photometer according to claim 7, wherein the preprocessing unit applies at least one of a rotation method and a double rotation method to the two signals from the light receiving unit. A light emitting unit that irradiates a living tissue with light of two different wavelengths;
In a pulse photometer comprising a light receiving unit that generates a signal in response to light generated from the light emitting unit and transmitted or reflected through a living tissue,
A frequency spectrum conversion step of obtaining frequencies and amplitudes at all peaks of the frequency spectrum from a frequency spectrum obtained by frequency analysis of the signal measured in a predetermined time zone;
A frequency correlation calculating step of calculating a frequency correlation for each peak from a difference between the frequency at all the peaks of the frequency spectrum and a previous pulse frequency that is a pulse frequency immediately before the predetermined time zone;
An amplitude correlation calculating step of calculating an amplitude correlation for each peak from the amplitude at all the peaks of the frequency spectrum;
A comprehensive correlation calculating step of calculating a total correlation for each peak from the frequency correlation for each peak and the amplitude correlation for each peak;
A pulse rate specifying method comprising a pulse rate specifying step of specifying a pulse rate in the predetermined time zone based on the total correlation degree, wherein the steps are executed in order.   The pulse rate specifying method according to claim 9, wherein, in the pulse rate specifying step, the frequency of the peak at which the total correlation is maximized is specified as the pulse rate in the predetermined time zone.   11. The pulse rate specifying method according to claim 9, wherein the pulse photometer sets the frequency of the specified pulse rate as the next previous pulse frequency. 11. In the frequency correlation calculation step,
Obtain the absolute value of the difference between the frequency and the previous pulse frequency at all peaks of the frequency spectrum,
Normalizing the absolute value of the difference so that the maximum absolute value of the difference is 1,
The pulse rate specifying method according to claim 9, wherein a value obtained by subtracting the absolute value of the difference normalized from 1 is used as the frequency correlation degree.   The amplitude correlation degree is a value obtained by multiplying a value obtained by normalizing the amplitude so that a maximum value of the amplitude becomes 1 and a weighting coefficient in the amplitude correlation degree calculating step. 12. The pulse rate identification method according to any one of 12 above.   The pulse rate specifying method according to claim 9, wherein the total correlation is a sum of a square of the frequency correlation and a square of the amplitude correlation.   15. The pre-processing step of reducing noise from the two signals from the light receiving unit corresponding to light having two different wavelengths is performed before the frequency spectrum conversion step. The pulse rate identification method of crab.   16. The pulse rate specifying method according to claim 15, wherein in the pre-processing step, at least one of a rotation method and a double rotation method is applied to the two signals from the light receiving unit.

Images