JP2024005557A - Heartbeat interval measuring apparatus, computer program for heartbeat interval measuring apparatus, and learning model generation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately extract a heartbeat peak from cardiac sound waveform data and accurately measure a heartbeat interval.

SOLUTION: A heartbeat interval measuring apparatus 10 comprises: a data acquisition unit 12 configured to acquire cardiac sound waveform data and biological data different from the cardiac sound waveform data; a heartbeat peak selection unit 14 configured to select a heartbeat peak of the cardiac sound waveform data on the basis of a feature amount of the biological data; and a heartbeat interval calculation unit 16 configured to calculate a heartbeat interval from the selected heartbeat peak.

SELECTED DRAWING: Figure 1

COPYRIGHT: (C)2024,JPO&INPIT

Description

The technology disclosed in this specification relates to a heartbeat interval measuring device, a computer program for the heartbeat interval measuring device, and a method for generating a learning model.

Patent Document 1 and Patent Document 2 disclose techniques for monitoring a user's psychological or physical condition from variations in heartbeat intervals.

Japanese Patent Application Publication No. 2017-124014 JP2020-075136A

By measuring the heartbeat interval using heart waveform data, the heartbeat interval can be measured with a simple device. In order to measure heartbeat intervals using heart waveform data, it is necessary to accurately extract heartbeat peaks from the heartbeat waveform data and calculate heartbeat intervals from the heartbeat peaks. The present specification provides a technique that can accurately measure heartbeat intervals by accurately extracting heartbeat peaks from heart waveform data.

A heartbeat interval measuring device (10) disclosed in this specification includes a data acquisition unit (12) configured to acquire heart waveform data and biological data different from the heart waveform data; A heartbeat peak selection unit (14) configured to select a heartbeat peak of the heartbeat waveform data based on the feature amount, and a heartbeat interval calculation unit (14) configured to calculate a heartbeat interval from the selected heartbeat peak. (16). According to this heartbeat interval measuring device, since the biological data when measuring the heartbeat waveform data is taken into account, the heartbeat peak of the heartbeat waveform data can be selected more accurately. Therefore, the heartbeat interval measuring device disclosed in this specification can accurately measure heartbeat intervals.

A computer program for a heartbeat interval measuring device (10) disclosed in this specification causes the heartbeat interval measuring device to acquire the following parts, that is, heart waveform data and biological data different from the heart waveform data. a data acquisition unit (12) configured to select a heartbeat peak in the heart waveform data based on the feature amount of the biological data; It may function as a heartbeat interval calculation section (16) configured to calculate heartbeat intervals from peaks. According to this computer program, since the biological data when the heart waveform data is measured is taken into consideration, the heartbeat peak of the heart waveform data can be selected more accurately. Therefore, the computer program for the heartbeat interval measuring device disclosed in this specification can accurately measure heartbeat intervals.

The method for generating a learning model disclosed in this specification acquires a plurality of training data including biological data, heart waveform data, and heartbeat peaks, inputs the biological data and heartbeat waveform data, and calculates the heartbeat peaks of the heartbeat waveform data. A learning model may be generated as an output. When the trained model generated in this manner is used, the biological data obtained when the heart waveform data is measured is taken into consideration, so that the heartbeat peak of the heart waveform data can be selected more accurately. Therefore, by using the learned model generated by the generation method disclosed in this specification, it is possible to accurately measure heartbeat intervals.

1 schematically shows a functional block diagram of a heartbeat interval measurement system. An example of respiratory waveform data and heart waveform data in one cycle of breathing is shown. 5 shows a processing flow of a method for selecting a heartbeat peak of heart waveform data according to a predetermined rule determined based on the feature amount of respiratory waveform data. A diagram for explaining the method of FIG. 3 is shown. A processing flow of a method of selecting a heartbeat peak of heart waveform data using sample heart waveform data is shown. An overview of the trained model is shown. The processing flow of a method for selecting heartbeat peaks in heart waveform data using a trained model is shown.

Hereinafter, a system for measuring a user's heartbeat interval will be described with reference to the drawings. By monitoring variations in the user's heartbeat interval measured by this system, the user's psychological or physical state can be estimated. For example, a user's drowsiness can be estimated by monitoring changes in the user's heartbeat interval. Such a system may be installed in a vehicle and used to detect driver drowsiness.

As shown in FIG. 1, the heartbeat interval measurement system 1 includes a heartbeat interval measurement device 10 and a sensor 20. The heartbeat interval measuring device 10 is configured by a computer, and includes a data acquisition section 12, a heartbeat peak selection section 14, and a heartbeat interval calculation section 16. The computer has a CPU, a storage device, and an input/output device, and by executing a program stored in the storage device on the CPU, the CPU is used as a data acquisition unit 12, a heartbeat peak selection unit 14, and a heartbeat interval calculation unit. 16, and can execute various processes described below. The sensor 20 is a sensor capable of detecting the user's heart sounds, and may be, for example, a microphone, although it is not particularly limited. The sensor 20, which is a microphone, inputs sounds emitted by the user (sounds including both heart sounds and breathing sounds, as described later), converts them into sensor signals, which are electrical signals, and outputs the sensor signals to the heartbeat interval measuring device 10.

The data acquisition unit 12 of the heartbeat interval measuring device 10 receives the sensor signal output from the sensor 20. The data acquisition unit 12 further extracts cardiac waveform data and respiratory waveform data from the received sensor signal. In this example, since the sensor 20 is configured with one microphone, the data acquisition unit 12 extracts heart waveform data and respiratory waveform data from the sensor signal based on the frequency band. Instead of this example, the sensor 20 may be configured with two sensors, ie, a sensor that selectively acquires cardiac waveform data and a sensor that selectively acquires respiratory waveform data. In this case, the data acquisition unit 12 receives heart waveform data from one sensor and receives respiratory waveform data from the other sensor.

FIG. 2 illustrates extracted cardiac waveform data and respiratory waveform data. (A) in FIG. 2 is respiratory waveform data, and (B) in FIG. 2 is heart waveform data. As shown in FIG. 2, the data acquisition unit 12 extracts respiratory waveform data from the sensor signal using a bandpass filter that passes a low frequency range, and extracts respiratory waveform data using a bandpass filter that passes a high frequency range. Extract heart waveform data from the sensor signal. In this example, heart waveform data for six beats appears in one cycle of breathing.

When the vertical axis in FIG. 2A is positive, it indicates that the breathing is in an "exhale" state. When the vertical axis in FIG. 2A is negative, it indicates that the breathing is in a "sucking" state. Each of T1 to T6 in FIG. 2 indicates, for one heartbeat, respiratory waveform data and heart waveform data that correspond in time series to that heartbeat. "T1" indicates respiratory waveform data and heart waveform data for one beat immediately after the "sucking" of breathing becomes the strongest. “T2” indicates one beat of respiratory waveform data and heart waveform data when breathing changes from “inhalation” to “exhalation”. "T3" shows the respiratory waveform data and heart waveform data when the exhalation of breathing becomes the strongest. "T4" shows the respiratory waveform data and heart waveform data immediately after the exhalation becomes the strongest. “T5” indicates one beat of respiratory waveform data and heart waveform data when breathing changes from “exhalation” to “inhalation”. "T6" indicates the respiratory waveform data and heart waveform data for one beat when the breathing force becomes the strongest. As shown at T1 to T6 in FIG. 2, the heart waveform data for one beat has various shapes depending on the state of breathing at that time. This suggests that the heartbeat peak of the heart waveform data to be selected for measuring the heartbeat interval can be selected more accurately by considering the state of respiration. As explained below, the heartbeat interval measurement system 1 is configured to select the heartbeat peak of the heart waveform data in consideration of the state of breathing.

The heartbeat peak selection unit 14 of the heartbeat interval measuring device 10 receives respiratory waveform data and heart waveform data from the data acquisition unit 12. The heartbeat peak selection unit 14 further selects a heartbeat peak for each heartbeat waveform data included in the heartbeat waveform data. More specifically, when selecting the heartbeat peak of one beat's worth of heartbeat waveform data, the heartbeat peak selection unit 14 selects respiratory waveform data that is chronologically associated with the heartbeat waveform data of one beat. The feature amount is calculated, and the heartbeat peak of the heart waveform data for that one beat is selected based on the feature amount. The feature amount of the respiratory waveform data is not particularly limited, but may be, for example, the slope or phase of the respiratory waveform.

As described above, the heart waveform data for one beat has various shapes depending on the state of breathing at that time. Therefore, the timing at which the heartbeat peak of one beat's worth of heartbeat waveform data should be determined is affected by the state of breathing at that time. Since the heartbeat peak selection unit 14 can select the heartbeat peak of one beat's worth of heartbeat waveform data based on the feature amount of the respiratory waveform data, the heartbeat peak can be selected more accurately.

The heartbeat interval calculation section 16 of the heartbeat interval measuring device 10 receives the heartbeat peak of the heartbeat waveform data from the heartbeat peak selection section 14 . The heartbeat interval calculation unit 16 further calculates the interval between chronologically adjacent heartbeat peaks, that is, the heartbeat interval. Since accurate heartbeat peaks are extracted, the calculated heartbeat intervals are also accurate. The heartbeat interval measurement system 1 can estimate the psychological or physical state of the user by monitoring fluctuations in the calculated heartbeat interval.

Below, some specific examples of selecting heartbeat peaks of heart waveform data from feature quantities of respiratory waveform data will be illustrated. The several specific examples to be described are just examples, and by employing various other techniques, the heartbeat peak of the cardiac waveform data can be selected from the feature amount of the respiratory waveform data. Furthermore, although an example will be described below in which heartbeat intervals are calculated using real-time processing, instead of this example, heartbeat intervals may be calculated using batch processing.

(Selection of heart rate peak according to predetermined rules)
FIG. 3 shows a processing flow of a method for selecting a heartbeat peak of heart waveform data according to a predetermined rule determined based on the feature amount of respiratory waveform data.

First, in step S1, the data acquisition unit 12 receives a sensor signal to be measured from the sensor 20.

Next, in step S2, the data acquisition unit 12 extracts cardiac waveform data and respiratory waveform data from the sensor signal based on the frequency band.

Next, in step S3, the heartbeat peak selection unit 14 selects the heartbeat peak of the heartbeat waveform data according to a rule determined based on the feature amount of the respiratory waveform data. In this example, as shown in FIG. 4, when selecting the heartbeat peak of one beat's worth of heartbeat waveform data, the heartbeat peak selection unit 14 chronologically selects the heartbeat peak of one beat's worth of heartbeat waveform data. If the associated respiratory waveform data is negative, the maximum value (indicated by a circle) of the heart waveform data is selected as the heartbeat peak. Furthermore, when selecting the heartbeat peak of one beat's worth of heartbeat waveform data, the heartbeat peak selection unit 14 further selects a heartbeat peak if the respiratory waveform data chronologically associated with the heartbeat waveform data of one beat is positive. , the maximum value (indicated by a triangle) at the next timing of the minimum value (indicated by a square) of the heart waveform data is selected as the heartbeat peak. In this way, the heartbeat peak selection unit 14 selects the heartbeat peak of the heartbeat waveform data according to a predetermined rule that sets a predetermined peak in the heartbeat waveform as a heartbeat peak, based on the sign (i.e., phase) of the respiratory waveform data. Select. Such a rule is an example. Rules for selecting heartbeat peaks may be set as appropriate, for example, depending on the environment in which the heartbeat interval measurement system 1 is used and/or the user who uses the heartbeat interval measurement system 1.

In step S4, the heartbeat interval calculation unit 16 calculates the heartbeat interval from the selected heartbeat peak.

In step S5, if the measurement is to be continued, the process returns to step S1; if the measurement is not to be continued, the process is finished.

According to this method, the heartbeat peak selection unit 14 selects the heartbeat peak of the heartbeat waveform data according to a predetermined rule determined based on the sign of the respiratory waveform data, that is, the phase of the respiratory waveform data. Such rules may be set by trial and error depending on the environment in which the heartbeat interval measurement system 1 is used (including, for example, the intended use). In this way, the heartbeat peak selection unit 14 can more accurately select the heartbeat peak of the heart waveform data by considering the state of breathing. Therefore, the heartbeat interval measuring device 10 can accurately measure heartbeat intervals.

(Selection of heartbeat peak using sample heart waveform data)
FIG. 5 shows a processing flow of a method for selecting a heartbeat peak of heart waveform data using sample heart waveform data.

First, in step S11, the data acquisition unit 12 receives a sample sensor signal from the sensor 20.

Next, in step S12, the data acquisition unit 12 extracts cardiac waveform data and respiratory waveform data from the sample sensor signal based on the frequency band.

Next, in step S13, the heartbeat peak selection unit 14 records a plurality of types of sample heartbeat waveform data chronologically associated with the respiratory waveform data. More specifically, the heartbeat peak selection unit 14 records a plurality of types of sample heart waveform data that are chronologically associated with characteristic breathing states in one cycle of breathing. In this example, as shown in FIG. 2, the heartbeat peak selection unit 14 records sample heart waveform data for six beats indicated by "T1" to "T6" of one cycle of respiration. The heartbeat peak selection unit 14 may sample only one cycle of respiration and record sample heart waveform data for multiple beats included therein, or may sample multiple cycles of respiration and select an appropriate one from among them. Sample heart waveform data for multiple beats may be recorded. Further, the heartbeat peak selection unit 14 may record sample heart waveform data for less than six beats, or may record sample heart waveform data for more than six beats. In this way, in this example, a plurality of types of sample heart waveform data are prepared from sample sensor signals prior to measurement. Instead of this example, a plurality of standard types of sample heart waveform data that can be universally applied to any user may be prepared in advance.

Next, in step S14, the data acquisition unit 12 receives a sensor signal to be measured from the sensor 20.

Next, in step S15, the data acquisition unit 12 extracts cardiac waveform data and respiratory waveform data from the sensor signal to be measured based on the frequency band.

Next, in step S16, the heartbeat peak selection unit 14 selects corresponding sample heart waveform data from a plurality of types of sample heart waveform data based on the feature amount (inclination or phase of the respiratory waveform) of the respiratory waveform data to be measured. select. Specifically, when selecting the heartbeat peak of the heartbeat waveform data for one beat of the measurement target, the heartbeat peak selection unit 14 selects the heartbeat peak that is chronologically associated with the heartbeat waveform data for one beat. Calculate the feature amount of the respiratory waveform data. The heartbeat peak selection unit 14 further selects sample heart waveform data associated with sample respiratory waveform data having a feature value closest to the calculated feature value. Referring to the example of FIG. 2, if the slope of the respiratory waveform to be measured is positive and large, the heartbeat peak selection unit 14 selects a sample heart sound corresponding to "T2" having a feature quantity closest to that feature quantity. Select waveform data.

Next, in step S17, the heartbeat peak selection unit 14 selects the peak of the correlation function between the selected sample heart waveform data and the heart waveform data to be measured as the heartbeat peak of the heartbeat waveform data to be measured.

In step S18, the heartbeat interval calculation unit 16 calculates the heartbeat interval from the selected heartbeat peak.

In step S19, if the measurement is to be continued, the process returns to step S14, and if the measurement is not to be continued, the process is finished.

According to this method, the heartbeat peak selection unit 14 selects the peak of the correlation function between the sample heart waveform data and the heart waveform data to be measured as the heartbeat peak of the heartbeat waveform data to be measured. Since the sample heart waveform data and the heartbeat waveform data to be measured are data obtained when the breathing conditions are similar, the peak of the correlation function can more accurately reflect the heartbeat peak. In this way, the heartbeat peak selection unit 14 can more accurately select the heartbeat peak of the heart waveform data by considering the state of breathing. Therefore, the heartbeat interval measuring device 10 can accurately measure heartbeat intervals.

The same plural types of sample heart waveform data in the above example were used as long as the measurement continued. Instead of this example, if the measurement is continued, the heartbeat peak selection unit 14 may accumulate the heartbeat waveform data to be measured as sample heartbeat waveform data when returning from step S19 to step S14. By accumulating sample heart waveform data, more segmented sample heart waveform data can be prepared, so that the heartbeat peak of the heartbeat waveform data can be selected more accurately. Further, if the measurement is continued, the heartbeat peak selection unit 14 may update the heartbeat waveform data to be measured as sample heartbeat waveform data when returning from step S19 to step S14. The update may be performed in units of one breath cycle. Further, the update may be executed for each cycle of breathing, or may be executed periodically after a plurality of cycles of breathing have elapsed. By updating the sample heart waveform data, the latest sample heart waveform data can be used and changes in the status of the user and the device can be followed, so the heartbeat peak of the heartbeat waveform data can be selected more accurately. Can be done.

(Selection of heart rate peak using trained model)
FIG. 6 schematically shows a trained model used by the heartbeat peak selection unit 14 of the heartbeat interval measuring device 10. This trained model is a machine learning model that inputs feature amounts of respiratory waveform data and heart waveform data, and outputs heartbeat peaks of the heart waveform data. The trained model may be generated, for example, by machine learning a neural network including an input layer, a hidden layer, and an output layer. The trained model may be a machine learning model that is generated using a deep learning technique and includes a plurality of hidden layers.

A trained model can be generated, for example, by applying a supervised learning method to a neural network. In this example, the feature amount of the respiratory waveform data, which is one of the inputs of the teacher data, includes at least one of the slope and phase of the respiratory waveform for one beat. Heart waveform data, which is one of the inputs of teacher data, includes heart waveform data for one beat. The heartbeat peak of the heart waveform data that is the output of the teacher data is set as the true value. The generation of the machine learning model may be executed by the heartbeat interval measuring device 10 (see FIG. 1), or may be executed by a device different from the heartbeat interval measuring device 10. The learned model is stored in the storage section of the heartbeat interval measuring device 10.

FIG. 7 shows a processing flow of a method for selecting heartbeat peaks in heart waveform data using a learned model.

First, in step S21, the data acquisition unit 12 receives a sensor signal to be measured from the sensor 20.

Next, in step S22, the data acquisition unit 12 extracts cardiac waveform data and respiratory waveform data from the sensor signal based on the frequency band.

Next, in step S23, the heartbeat peak selection unit 14 inputs the feature amount of the respiratory waveform data and the heartbeat waveform data to the trained model, and selects the heartbeat peak of the heartbeat waveform data.

In step S24, the heartbeat interval calculation unit 16 calculates the heartbeat interval from the selected heartbeat peak.

In step S25, if the measurement is to be continued, the process returns to step S21, and if the measurement is not to be continued, the process is finished.

According to this method, the heartbeat peak selection unit 14 selects the heartbeat peak of the heart waveform data using the learned model. The learned model selects the heartbeat peak of the heart waveform data in consideration of the feature amount of the respiratory waveform data. In this way, the heartbeat peak selection unit 14 can more accurately select the heartbeat peak of the heart waveform data by considering the state of breathing. Therefore, the heartbeat interval measuring device 10 can accurately measure heartbeat intervals.

In each of the above examples, an example has been described in which respiratory waveform data is used as biological data to measure heartbeat intervals. This example is just an example, and other biological data, such as data indicating the user's posture, the user's voice, and vibration data generated from within the user's body while chewing a drink, etc., may also be used.

The features of the technology disclosed in this specification will be summarized below. The technical elements described below are independent technical elements that exhibit technical utility alone or in various combinations, and are not limited to the combinations described in the claims as filed. do not have.

(Feature 1)
The heart rate interval measuring device includes a data acquisition unit configured to acquire heart sound waveform data and biological data different from the heart sound waveform data; A heartbeat interval measuring device comprising: a heartbeat peak selection section configured to select a heartbeat peak of waveform data; and a heartbeat interval calculation section configured to calculate a heartbeat interval from the selected heartbeat peak. .

(Feature 2)
The heartbeat interval measuring device according to feature 1, wherein the biological data is respiratory waveform data.

(Feature 3)
The heartbeat interval measuring device according to feature 2, wherein the data acquisition unit is configured to extract cardiac waveform data and respiratory waveform data from sensor signals acquired from the same sensor based on a frequency band.

(Feature 4)
The heartbeat interval measuring device according to feature 2 or 3, wherein the feature amount of the respiratory waveform data is the slope or phase of the respiratory waveform.

(Feature 5)
According to any one of features 1 to 4, the heartbeat peak selection unit is configured to select the heartbeat peak of the heartbeat waveform data according to a predetermined rule determined based on the feature amount of the biometric data. Heartbeat interval measuring device.

(Feature 6)
The heartbeat peak selection unit prepares multiple types of sample heart waveform data chronologically associated with biological data, and based on the feature amount of the biological data when measuring the heart waveform data to be measured. selecting corresponding sample heart waveform data from a plurality of types of sample heart waveform data, and selecting the peak of a correlation function between the selected sample heart waveform data and the heart waveform data to be measured as the heartbeat peak of the heartbeat waveform data to be measured; The heartbeat interval measuring device according to any one of features 1 to 4, which is configured as follows.

(Feature 7)
The heartbeat interval measuring device according to feature 6, wherein the heartbeat peak selection section is further configured to accumulate and/or update the heartbeat waveform data to be measured as sample heartbeat waveform data.

(Feature 8)
The heartbeat peak selection unit inputs biological data and heart waveform data, and applies the biological data and heart waveform data to be measured to a trained model trained using teacher data that outputs heartbeat peaks of the heartbeat waveform data. The heartbeat interval measuring device according to any one of features 1 to 4, wherein the heartbeat interval measuring device is configured to select a heartbeat peak of the heart waveform data by inputting the following.

(Feature 9)
The heartbeat interval measuring device according to feature 1, wherein the biometric data is the user's posture.

(Feature 10)
A computer program for a heartbeat interval measuring device, the computer program being configured to acquire the following parts of the heartbeat interval measuring device, namely, heart waveform data and biological data different from the heart waveform data. an acquisition unit; a heartbeat peak selection unit configured to select a heartbeat peak of the heart waveform data based on the feature amount of the biological data; and a heartbeat peak selection unit configured to calculate a heartbeat interval from the selected heartbeat peak. A computer program configured to function as a heartbeat interval calculation unit.

(Feature 11)
A method for generating a learning model, in which multiple training data including biological data, heart waveform data, and heartbeat peaks are acquired, the biological data and heartbeat waveform data are input, and the heartbeat peak of the heartbeat waveform data is output. How to generate the model.

Although the embodiments have been described in detail above, these are merely examples and do not limit the scope of the claims. The techniques described in the claims include various modifications and changes to the specific examples illustrated above. The technical elements described in this specification or the drawings exhibit technical usefulness singly or in various combinations, and are not limited to the combinations described in the claims as filed. Furthermore, the techniques illustrated in this specification or the drawings simultaneously achieve multiple objectives, and achieving one of the objectives has technical utility in itself.

1: Heartbeat interval measurement system, 10: Heartbeat interval measurement device, 12: Data acquisition unit, 14: Heartbeat peak selection unit, 16: Heartbeat interval calculation unit, 20: Sensor

Claims (11)

A heartbeat interval measuring device (10),
a data acquisition unit (12) configured to acquire heart waveform data and biological data different from the heart waveform data;
a heartbeat peak selection unit (14) configured to select a heartbeat peak of the heartbeat waveform data based on the feature amount of the biological data;
A heartbeat interval measuring device, comprising: a heartbeat interval calculating section (16) configured to calculate a heartbeat interval from the selected heartbeat peak. The heartbeat interval measuring device according to claim 1, wherein the biological data is respiratory waveform data. The data acquisition unit includes:
The heartbeat interval measuring device according to claim 2, configured to extract the cardiac waveform data and the respiratory waveform data from sensor signals acquired from the same sensor based on a frequency band. The heartbeat interval measuring device according to claim 2, wherein the feature amount of the respiratory waveform data is a slope or a phase of the respiratory waveform. The heartbeat peak selection section includes:
The heartbeat interval according to any one of claims 1 to 4, wherein the heartbeat peak of the heart waveform data is selected according to a predetermined rule determined based on the feature amount of the biological data. measuring device. The heartbeat peak selection section includes:
preparing multiple types of sample heart waveform data chronologically associated with the biological data;
selecting corresponding sample heart waveform data from the plurality of types of sample heart waveform data based on the feature amount of the biological data when the heart waveform data of the measurement target is measured;
Any one of claims 1 to 4, wherein the peak of a correlation function between the selected sample heart waveform data and the heart waveform data to be measured is selected as the heartbeat peak of the heartbeat waveform data to be measured. The heartbeat interval measuring device according to item 1. The heartbeat peak selection section further includes:
The heartbeat interval measuring device according to claim 6, configured to accumulate and/or update the heart waveform data to be measured as the sample heart waveform data. The heartbeat peak selection section includes:
The biological data and the cardiac sound waveform data to be measured are input to a trained model that is trained using teacher data in which the biological data and the cardiac sound waveform data are input and the heartbeat peak of the cardiac soundwave data is output. The heartbeat interval measuring device according to any one of claims 1 to 4, wherein the heartbeat interval measuring device is configured to select the heartbeat peak of the heartbeat waveform data by inputting the following. The heartbeat interval measuring device according to claim 1, wherein the biometric data is a user's posture. A computer program for a heartbeat interval measuring device (10), which comprises the following parts:
a data acquisition unit (12) configured to acquire heart waveform data and biological data different from the heart waveform data;
a heartbeat peak selection unit (14) configured to select a heartbeat peak of the heartbeat waveform data based on the feature amount of the biological data;
A computer program that functions as a heartbeat interval calculation section (16) configured to calculate a heartbeat interval from the selected heartbeat peak. A method of generating a learning model, the method comprising:
A method of acquiring a plurality of teacher data including biological data, heart waveform data, and heartbeat peaks, inputting the biological data and the heartbeat waveform data, and generating a learning model whose output is the heartbeat peak of the heartbeat waveform data.

Images