Measurement quality evaluation device, method, and program

Abstract

A first arithmetic circuit (101) detects R waves from an electrocardiographic waveform (electrocardiogram) measured by an electrocardiograph (121) to determine an interval between the adjacent R waves (heartbeat interval: RRI). A second arithmetic circuit (102) determines, from the interval determined by the first arithmetic circuit (101), an instantaneous heart rate for each pulsation indicated by the R wave. A third arithmetic circuit (103) determines a smooth heart rate obtained by smoothing the instantaneous heart rates of respective pulsations calculated by the second arithmetic circuit (102). A fourth arithmetic circuit (104) determines a signal-to-noise ratio of the electrocardiographic waveform using the instantaneous heart rate and the smooth heart rate. A transmission circuit (111) transmits the signal-to-noise ratio determined by the fourth arithmetic circuit (104) to a preset transmission destination.

Description

[Technical field]

[0001] The present invention relates to an apparatus, method, and program for evaluating measurement quality for determining a signal-to-noise ratio in an electrocardiogram. [Background technology]

[0002] In recent years, a technology has been proposed for detecting electrocardiogram information of a user (subject) from a sensor worn by the user (subject) (Non-Patent Document 1). In such a technology, since the electrocardiogram information generally contains artifacts caused by the user's body movements, the measurement quality, i.e., the clarity of the electrocardiogram shape, may be reduced. In response to such a reduction in measurement quality, a technology such as that in Patent Document 1 has been proposed as a method for evaluating the measurement quality based on the validity of the heart rate value calculated from the electrocardiogram. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2018-11819 A [Non-patent literature]

[0004] [Non-Patent Document 1] Nippon Telegraph and Telephone Corporation, "Development of low-power, small-sized wearable sensor that enables measurement of electrocardiogram, acceleration, temperature and humidity for smart healthcare," NTT Holding Company News Release, November 8, 2019, [Retrieved September 2, 2021], (https: / / www.ntt.co.jp / news2019 / 1911 / 191108a.html). Summary of the Invention [Problem to be solved by the invention]

[0005] When evaluating the quality of measurement in this way, the technology of Patent Document 1 evaluates the validity of the numerical value of the heart rate, so the quality of the original electrocardiogram remains unknown. Therefore, since it is not possible to obtain information that may have contributed to the deterioration of the quality of the electrocardiogram, such as the size of the artifact that was mixed into the electrocardiogram, there is a problem that the user cannot take any corrective action that leads to a fundamental solution.

[0006] The present invention has been made to solve the above problems, and has an object to provide a high-quality electrocardiogram. [Means for solving the problem]

[0007] The measurement quality evaluation device according to the present invention includes a first arithmetic circuit that detects R waves from an electrocardiogram waveform measured by an electrocardiograph and determines the interval between adjacent R waves; a second arithmetic circuit that determines an instantaneous heart rate for each beat indicated by an R wave from the interval determined by the first arithmetic circuit; a third arithmetic circuit that determines a smoothed heart rate by smoothing the instantaneous heart rate for each beat calculated by the second arithmetic circuit; a fourth arithmetic circuit that determines a signal-to-noise ratio of the electrocardiogram waveform using at least one of the instantaneous heart rate and the smoothed heart rate; and a transmission circuit that transmits the signal-to-noise ratio determined by the fourth arithmetic circuit to a set destination.

[0008] The measurement quality evaluation method of the present invention includes a first step of detecting R waves from an electrocardiogram waveform measured by an electrocardiograph and determining the interval between adjacent R waves, a second step of determining the instantaneous heart rate for each beat indicated by the R wave from the interval determined in the first step, a third step of determining a smoothed heart rate by smoothing the instantaneous heart rate for each beat calculated in the second step, a fourth step of determining a signal-to-noise ratio of the electrocardiogram waveform using the instantaneous heart rate and the smoothed heart rate, and a fifth step of transmitting the signal-to-noise ratio determined in the fourth step to a set destination.

[0009] A program according to the present invention is a program for causing a computer to execute the above-mentioned measurement quality evaluation method. Effect of the Invention

[0010] As described above, according to the present invention, the signal-to-noise ratio of an electrocardiogram waveform is obtained using the instantaneous heart rate and the smoothed heart rate, so that a high-quality electrocardiogram can be provided. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 is a configuration diagram showing a configuration of a measurement quality evaluation device according to a first embodiment of the present invention. [Diagram 2] FIG. 2 is a flowchart illustrating the measurement quality evaluation method according to the first embodiment of the present invention. [Figure 3A] FIG. 3A is a correlation diagram showing the correlation between the instantaneous heart rate, the smoothed heart rate, and the SNR obtained (estimated) from the learning results of machine learning using the SNR, and the SNR in the training data of the SNR used in the machine learning. [Figure 3B] FIG. 3B is a correlation diagram showing the correlation between the instantaneous heart rate, smoothed heart rate, and the SNR obtained (estimated) from the learning results of machine learning using the SNR, and the SNR in the training data of the SNR used in the machine learning. [Figure 4] FIG. 4 is a diagram showing the hardware configuration of the measurement quality evaluation device. [Diagram 5] FIG. 5 is a configuration diagram for explaining a system using the measurement quality evaluation device in the first embodiment. [Figure 6] FIG. 6 is a configuration diagram showing the configuration of a system using the measurement quality evaluation device in the first embodiment. [Figure 7] FIG. 7 is a configuration diagram showing the configuration of a measurement quality evaluation device according to the second embodiment of the present invention. [Figure 8] FIG. 8 is a characteristic diagram showing an example of the determination information used in the measurement quality evaluation method according to the second embodiment. [Figure 9] FIG. 9 is a configuration diagram showing the configuration of a measurement quality evaluation device according to the third embodiment of the present invention. [Figure 10] FIG. 10 is a configuration diagram showing the configuration of a measurement quality evaluation device according to the fifth embodiment of the present invention. [Figure 11A]FIG. 11A is a correlation diagram showing the relationship between SNR and RRI when SNR>24. [Figure 11B] FIG. 11B is a correlation diagram showing the relationship between SNR and RRI when SNR=12. [Figure 11C] FIG. 11C is a correlation diagram showing the relationship between SNR and RRI when SNR=−6. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, a measurement quality evaluation device according to an embodiment of the present invention will be described.

[0013] [Embodiment 1] First, a measurement quality evaluation device according to a first embodiment of the present invention will be described with reference to Fig. 1. This measurement quality evaluation device includes a first arithmetic circuit 101, a second arithmetic circuit 102, a third arithmetic circuit 103, a fourth arithmetic circuit 104, and a transmission circuit 111.

[0014] The first arithmetic circuit 101 detects R waves from an electrocardiogram waveform (electrocardiogram) measured by the electrocardiograph 121, and calculates the interval between adjacent R waves (heart beat interval: RRI). The second arithmetic circuit 102 calculates an instantaneous heart rate for each beat indicated by an R wave from the interval calculated by the first arithmetic circuit 101. The third arithmetic circuit 103 calculates a smoothed heart rate by smoothing the instantaneous heart rate for each beat calculated by the second arithmetic circuit 102.

[0015] The fourth arithmetic circuit 104 calculates a signal-to-noise ratio of the electrocardiogram waveform using at least one of the instantaneous heart rate and the smoothed heart rate. The transmission circuit 111 transmits the signal-to-noise ratio calculated by the fourth arithmetic circuit 104 to a preset destination.

[0016] This measurement quality evaluation device can also include a fifth arithmetic circuit 105. The fifth arithmetic circuit 105 obtains a period during which the signal-to-noise ratio obtained by the fourth arithmetic circuit 104 falls below a certain value within the measurement period during which the electrocardiogram information is measured, and displays this on the display device 112. The fifth arithmetic circuit 105 also obtains a ratio between this measurement period and a period during which the signal-to-noise ratio obtained by the fourth arithmetic circuit 104 falls below a certain value within the measurement period during which the electrocardiogram information is measured, and displays this on the display device 112.

[0017] Next, the measurement quality evaluation method (operation of the measurement quality evaluation device) according to the first embodiment of the present invention will be described with reference to FIG.

[0018] First, in a first step S101, the first arithmetic circuit 101 detects R waves from an electrocardiogram waveform measured by the electrocardiograph 121 and calculates the interval between adjacent R waves. For example, the first arithmetic circuit 101 can calculate the RRI from the electrocardiogram by using the peak detection described in Reference 1.

[0019] Next, in a second step S102, the second arithmetic circuit 102 calculates the instantaneous heart rate for each beat indicated by the R wave from the interval obtained in the first step S101. For example, a method for calculating the instantaneous heart rate from the RRI can be to divide 60,000 by the RRI to obtain the instantaneous heart rate. This is because the unit of the instantaneous heart rate is beats per minute (bpm), and therefore the instantaneous heart rate [bpm (beats / min)] = 60 [seconds / min] x 1,000 [milliseconds / sec] ÷ RRI [milliseconds / beat].

[0020] Next, in a third step S103, the third arithmetic circuit 103 calculates a smoothed heart rate by smoothing the instantaneous heart rate for each beat calculated in the second step S102. For example, the smoothed heart rate can be a moving average using the most recent instantaneous heart rate data. Also, the smoothed heart rate can be a value obtained by applying an FIR filter or an IIR filter to the most recent instantaneous heart rate.

[0021] Next, in a fourth step S104, the fourth arithmetic circuit 104 calculates the signal-to-noise ratio of the electrocardiogram waveform using the instantaneous heart rate and the smoothed heart rate.

[0022] As a method for calculating the signal-to-noise ratio (SNR), for example, machine learning can be used. Electrocardiogram data with few artifacts and noise data in which only artifacts are recorded are prepared in advance, and training data is created by combining them.

[0023] The noise data is combined with the electrocardiogram data at different amplitude values ​​to create data sets of various SNRs. For the electrocardiogram data, publicly available data disclosed in Reference 2 may be used. For the noise data, publicly available data disclosed in Reference 3 may be used. Reference 3 provides six types of data sets in which the electrocardiogram data and noise data in Reference 2 are combined to change the SNR from -6 to 24 dB as an example of creating data sets of various SNRs, and these may be used as they are as SNR data sets.

[0024] Figure 3A is a correlation diagram showing the correlation between the SNR obtained (estimated) by the learning results of machine learning using instantaneous heart rate, smoothed heart rate, and SNR, and the SNR in the training data of the SNR used in machine learning. There is a positive correlation between the SNR of the data set and the predicted SNR, with a coefficient of determination of 0.75 and an error (RMSE) of 7.83, which indicates that the SNR can be predicted extremely well. Figure 3A shows the results of learning using the electrocardiogram data from Reference 2 and the noise data from Reference 4, using Random Forest, a type of machine learning method.

[0025] In this way, the SNR of the electrocardiogram can be calculated by using the judgment information obtained in advance through machine learning using instantaneous heart rate, smooth heart rate, and various SNR data sets.

[0026] In addition, the machine learning can use only the data set of the instantaneous heart rate and SNR without using the smoothed heart rate. Figure 3B shows the learning result in this case. Figure 3B shows the result of learning using the electrocardiogram data from Reference 2 and the noise data from Reference 4, using Random Forest, one of the machine learning methods. The coefficient of determination is 0.79, which is as good as before, but the error (RMSE) is 8.00, which is slightly worse because the smoothed heart rate is not used. However, it has the advantage of being easier to implement because it reduces the number of variables used for input by one. In addition, machine learning is not limited to Random Forest, and support vector machines, neural networks, logistic regression, ensemble learning, etc. can be used. In addition, the signal-to-noise ratio of the electrocardiogram waveform can also be obtained by multiple regression analysis.

[0027] Next, in a fifth step S105, the transmission circuit 111 transmits the signal-to-noise ratio calculated in the fourth step S104 to the set destination. In addition, in a sixth step S106, the fifth arithmetic circuit 105 calculates a period during which the calculated signal-to-noise ratio falls below a certain value within the measurement period during which the electrocardiogram information is measured, and displays this on the display device 112. In addition, the fifth arithmetic circuit 105 calculates a ratio between this measurement period and the period during which the calculated signal-to-noise ratio falls below a certain value within the measurement period during which the electrocardiogram information is measured, and displays this on the display device 112.

[0028] In this way, the calculated SNR can be displayed to the user, for example, to indicate the period during which the SNR fell below a certain value within the measurement period during which the electrocardiogram information was measured, or the ratio of the measurement period during which the electrocardiogram information was measured to the period during which the SNR fell below a certain value, thereby providing the user with appropriate information about the electrocardiogram measurement quality.

[0029] As shown in Fig. 4, the above-mentioned measurement quality evaluation device may be a computer device including a CPU (Central Processing Unit) 301, a main memory device 302, an external memory device 303, a network connection device 304, etc., and the above-mentioned functions (measurement quality evaluation method) may be realized by the CPU 301 operating (executing) a program loaded in the main memory device 302. The above-mentioned program is a program for causing a computer to execute the measurement quality evaluation method shown in the above-mentioned embodiment. The network connection device 304 is connected to a network 305. Also, each function may be distributed to multiple computer devices.

[0030] The measurement quality evaluation device according to the above-mentioned embodiment can also be configured by a programmable logic device (PLD) such as a field-programmable gate array (FPGA). For example, the logic element of the FPGA can be provided with a memory unit, a first arithmetic circuit, a second arithmetic circuit, a third arithmetic circuit, a fourth arithmetic circuit, and a transmission circuit to function as a measurement quality evaluation device. Each of the memory circuit, the first arithmetic circuit, the second arithmetic circuit, the third arithmetic circuit, the fourth arithmetic circuit, and the transmission circuit can be written into the FPGA by connecting a predetermined writing device. Furthermore, each of the above circuits written into the FPGA can be confirmed by a writing device connected to the FPGA.

[0031] Next, a system using the measurement quality evaluation device in the first embodiment will be described. For example, as shown in Fig. 5, a sensor terminal 202 is attached to the trunk of a subject 201, and the results of measurement by an electrocardiograph built into the sensor terminal 202 are relayed by a relay terminal 203 and transmitted to an external terminal 204. The sensor terminal 202 can be a computer device such as a smartphone or a tablet. The relay terminal 203 and the external terminal 204 can be computer devices such as a server device.

[0032] 6, the sensor terminal 202 includes an electrocardiograph 121, a conversion circuit 122, a memory 123, an arithmetic circuit 124, a transmission processing circuit 125, and a communication interface 126. The relay terminal 203 includes a communication interface 131, a reception processing circuit 132, a memory 133, an arithmetic circuit 134, a transmission processing circuit 135, and a communication interface 136. The external terminal 204 includes a communication interface 141, a reception processing circuit 142, a memory 143, an arithmetic circuit 144, a control circuit 145, and an operating device 146.

[0033] The conversion circuit 122 converts the analog acceleration signal measured by the electrocardiograph 121 into digital data at a predetermined sampling rate and outputs the digital data. The memory 123 stores the electrocardiogram waveform digitized by the conversion circuit 122. The arithmetic circuit 124 calculates the RRI based on the electrocardiogram waveform stored in the memory 123. The arithmetic circuit 124 also calculates an instantaneous heart rate for each beat indicated by an R wave from the calculated RRI. The arithmetic circuit 124 also calculates a smoothed heart rate by smoothing the calculated instantaneous heart rate for each beat. The arithmetic circuit 124 also calculates a signal-to-noise ratio of the electrocardiogram waveform using the instantaneous heart rate and the smoothed heart rate. For example, the arithmetic circuit 124 can be a first arithmetic circuit 101, a second arithmetic circuit 102, a third arithmetic circuit 103, and a fourth arithmetic circuit 104.

[0034] The transmission processing circuit 125 transmits the processing result (e.g., RRI) processed by the arithmetic circuit 124 to the relay terminal 203 through the communication interface 126. The communication interface 126 is composed of an arithmetic interface and an antenna compatible with wireless data communication standards such as LTE (Long Term Evolution), a third generation mobile communication system, a wireless LAN (Local Area Network), and Bluetooth (registered trademark).

[0035] The relay terminal 203 is composed of a communication interface 131 that receives data transmitted from the sensor terminal 202, a receiving processing circuit 132, a memory 133, an arithmetic circuit 134, a transmitting processing circuit 135, and a communication interface 136 that transmits data to the external terminal 204.

[0036] The external terminal 204 comprises a communication interface 141 that receives data transmitted from the relay terminal 203, a receiving processing circuit 142, a memory 143, an arithmetic circuit 144, and a control circuit 145 that issues operating commands to an operating device 146 that operates based on the analyzed data.

[0037] Based on the information (electrocardiogram waveform and the signal-to-noise ratio of the electrocardiogram waveform) stored in the memory 143, the control circuit 145 causes the operation device 146 to perform an operation to support the person being measured.

[0038] The operating device 146 is a video output device (monitor, etc.), an audio output device (speaker, musical instrument, etc.), a light source (LED: Light Emitting Diode or light bulb), an actuator (transducer, robot arm, electrotherapy device), a thermal device (heater or Peltier element), etc.

[0039] The arithmetic circuit 124 does not need to include all of the first arithmetic circuit 101, the second arithmetic circuit 102, the third arithmetic circuit 103, and the fourth arithmetic circuit 104, but can be distributed to the arithmetic circuit 134 and the arithmetic circuit 144.

[0040] [Embodiment 2] Next, a measurement quality evaluation device according to a second embodiment of the present invention will be described with reference to Fig. 7. This measurement quality evaluation device comprises a first arithmetic circuit 101, a second arithmetic circuit 102, a third arithmetic circuit 103, a fourth arithmetic circuit 104a, a transmission circuit 111, a fifth arithmetic circuit 105, and a sixth arithmetic circuit 106. The first arithmetic circuit 101, the second arithmetic circuit 102, the third arithmetic circuit 103, the fifth arithmetic circuit 105, and the transmission circuit 111 are the same as those in the first embodiment described above.

[0041] The sixth arithmetic circuit 106 judges the appropriateness of the smoothed heart rate calculated by the third arithmetic circuit 103. The fourth arithmetic circuit 104a calculates the signal-to-noise ratio of the electrocardiogram waveform based on the result of the judgment by the sixth arithmetic circuit 106.

[0042] In the measurement quality evaluation method according to the second embodiment, in a first step, the first arithmetic circuit 101 detects R waves from an electrocardiogram waveform measured by the electrocardiograph 121 and calculates the interval between adjacent R waves. Next, in a second step, the second arithmetic circuit 102 calculates the instantaneous heart rate for each beat indicated by the R wave from the interval found in the first step.

[0043] Next, in a third step, the third arithmetic circuit 103 calculates a smoothed heart rate by smoothing the instantaneous heart rate for each beat calculated in the second step S102. Also, in a sixth step, the sixth arithmetic circuit 106 judges the appropriateness of the smoothed heart rate calculated in the third step. Next, in a fourth step, the fourth arithmetic circuit 104a calculates the signal-to-noise ratio of the electrocardiogram waveform based on the result of the judgment in the sixth step.

[0044] The judgment of the appropriateness of the smoothed heart rate can be realized, for example, by using the method described in claim 1 of Patent Document 1. The judgment information for calculating the SNR using the output from the sixth arithmetic circuit 106 (i.e., the judgment result of whether the smoothed heart rate is appropriate or inappropriate) and the above-mentioned SNR data set is generated in advance by machine learning, and the fourth arithmetic circuit 104a stores this information.

[0045] FIG. 8 shows an example of the judgment information. In FIG. 8, the vertical axis is SNR, and the horizontal axis is the ratio of results judged as appropriate to the judgment results of appropriateness. For example, if 60 judgment results are obtained per minute, and 50 of them are judged as appropriate, the value corresponding to the horizontal axis is 50 / 60 ≒ 0.83. In this case, according to the relationship in FIG. 8, the SNR is determined to be about 8 dB. If the value corresponding to the horizontal axis is less than 0.44 or greater than 0.87, the SNR cannot be determined from FIG. 8, but in these cases, an output that can be regarded as less than -24 dB or greater than 24 dB should be given.

[0046] In this way, by calculating the signal-to-noise ratio of the electrocardiogram waveform based on the judgment information previously learned by machine learning based on the appropriateness of the instantaneous heart rate, existing technologies such as Patent Document 1 can be effectively utilized, and it is possible to provide users with implementation options that provide appropriate electrocardiogram measurement quality. In addition, while two inputs were required for the fourth arithmetic circuit 104 in the first embodiment, only one input is required in the second embodiment, so that the configuration of the fourth arithmetic circuit 104a can be simplified.

[0047] [Embodiment 3] Next, a measurement quality evaluation device according to a third embodiment of the present invention will be described with reference to Fig. 9. This measurement quality evaluation device comprises a first arithmetic circuit 101, a second arithmetic circuit 102, a third arithmetic circuit 103, a fourth arithmetic circuit 104, a fifth arithmetic circuit 105, a transmission circuit 111a, and a seventh arithmetic circuit 107. The first arithmetic circuit 101, the second arithmetic circuit 102, the third arithmetic circuit 103, the fourth arithmetic circuit 104, and the fifth arithmetic circuit 105 are the same as those in the first embodiment described above.

[0048] The seventh arithmetic circuit 107 generates calculation impossible information when an R wave is not detected for a set period in the first arithmetic circuit 101. Furthermore, the transmission circuit 111a transmits the calculation impossible information.

[0049] In the measurement quality evaluation method according to the third embodiment, in a first step, the first arithmetic circuit 101 detects R waves from an electrocardiogram waveform measured by the electrocardiograph 121 and calculates the interval between adjacent R waves. Next, in a second step, the second arithmetic circuit 102 calculates the instantaneous heart rate for each beat indicated by the R wave from the interval found in the first step.

[0050] Next, in a third step, the third arithmetic circuit 103 calculates a smoothed heart rate by smoothing the instantaneous heart rate for each beat calculated in the second step S102. In addition, in a seventh step, the seventh arithmetic circuit 107 generates non-calculation information when an R wave is not detected during the period set in the first step. In a fourth step, the fourth arithmetic circuit 104 calculates the signal-to-noise ratio of the electrocardiogram waveform using the instantaneous heart rate and the smoothed heart rate. After this, in a fifth step, the transmission circuit 111 transmits the signal-to-noise ratio calculated in the fourth step and the non-calculation information generated in the seventh step to the set destination.

[0051] In the above-mentioned first embodiment, the SNR is calculated on the premise that the RRI is obtained as a value, so if the R wave cannot be detected and the RRI cannot be obtained, the SNR cannot be calculated. In order to deal with such a case, the seventh arithmetic circuit 107 generates calculation impossible information indicating that the RRI has not been detected when the RRI cannot be obtained within a certain period of time (for example, 2 seconds, which exceeds the pulsation period of a typical heartbeat). The calculation impossible information is notified as a substitute for not updating the SNR. This makes it possible to notify that a problem has occurred in the measurement, i.e., that the measurement quality is not good, even in a situation where the SNR cannot be obtained.

[0052] [Embodiment 4] Next, a measurement quality evaluation device according to a fourth embodiment of the present invention will be described. This measurement quality evaluation device is a modified example of the above-mentioned first to third embodiments. The fourth embodiment is characterized in that, when data from a sensor terminal is lost due to a communication interruption (packet loss), a reception processing circuit in a relay terminal or an external terminal notifies a memory of loss information.

[0053] If the SNR calculated in the above-described embodiment or the incalculable information determined in the third embodiment is lost as a packet loss when transferred to a relay terminal or an external terminal, the information will not reach the user. However, if the occurrence of packet loss is stored and notified, it is not necessary to conclude that there was a problem in the data detection or analysis process of the sensor terminal. This makes it possible to provide accurate information on the cause even in a situation where SNR information does not reach the user.

[0054] [Embodiment 5] Next, a measurement quality evaluation device according to a fifth embodiment of the present invention will be described with reference to Fig. 10. This measurement quality evaluation device includes a first arithmetic circuit 101, a fourth arithmetic circuit 104b, a fifth arithmetic circuit 105, a transmission circuit 111b, and an eighth arithmetic circuit 108. The first arithmetic circuit 101 and the fifth arithmetic circuit 105 are the same as those in the first embodiment described above.

[0055] The eighth arithmetic circuit calculates an SNR based on the RRI calculated by the first arithmetic circuit 101, or determines arrhythmia based on the RRI. The transmission circuit 111b transmits the SNR calculated by the eighth arithmetic circuit .

[0056] The relationship between SNR and RRI is shown in Figures 11A, 11B, and 11C. In the graph of SNR>24 in Figure 11A, the RRI fluctuates around 800 ms, but sometimes shows around 1000 ms or 500 ms, which is due to fluctuations caused by arrhythmia.

[0057] However, as the SNR decreases from SNR=12 in Fig. 11B to SNR=-6 in Fig. 11C, the number of points significantly deviating from the RRI of 800 ms increases. This is because the first arithmetic circuit 101 erroneously detects artifacts as R waves.

[0058] Focusing on this trend, the SNR can be grasped by using a threshold value. For example, when the RRI is in the range of 400 < RRI < 1200, it can be estimated that the "SNR is high"; when 1200 < RRI < 1700, the "SNR is medium"; and when RRI < 400 or 1700 < RRI, the "SNR is low". Also, it is possible to estimate whether the variation in the RRI value is due to arrhythmia. For example, when an RRI that deviates from the average value by 100 ms to 200 ms is obtained, if it is in the range of 400 < RRI < 1200, it can be determined that "the possibility of arrhythmia is high", and if RRI < 400 or 1200 < RRI, it can be determined that "the possibility of artifact is high", thus enabling the identification of arrhythmia and artifacts.

[0059] In addition, when the measured electrocardiogram waveform (electrocardiogram) is recorded over several days, an averaging processing circuit is provided to calculate an average time series using the RRI values obtained from the electrocardiograms at the same time zone each day. The averaging processing circuit calculates the average time series using only the instantaneous heart rate or only the smoothed heart rate when the SNR calculated in the fourth arithmetic circuit is above a certain level.

[0060] The above averaging processing circuit may use, for example, the technique of Reference 4. Thereby, only the highly reliable instantaneous heart rate or smoothed heart rate detected in a situation with a high SNR can be handled, and a highly reliable average time series can be calculated.

[0061] As described above, according to the present invention, since the signal-to-noise ratio of the electrocardiogram waveform is obtained using the instantaneous heart rate and the smoothed heart rate, a high-quality electrocardiogram can be provided. According to the present invention, the magnitude of the artifact mixed in the electrocardiogram waveform can be estimated, and a high-quality electrocardiogram waveform can be provided.

[0062] It should be noted that the present invention is not limited to the embodiments described above, and it is obvious that many modifications and combinations can be implemented by those with ordinary knowledge in the art within the technical idea of the present invention.

[0063] [References 1] Analog Devices, "Let's create a healthcare IoT that visualizes stress using an electrocardiogram sensor!", IoT course for web engineers taught by hardware professionals, No. 8, [Retrieved September 2, 2021], (https: / / www.analog.com / jp / education / landing-pages / 003 / iot_lectureship / content_08.html). [Reference 2] G. Moody and R. Mark, "MIT-BIH Arrhythmia Database", [Retrieved September 2, 2021], (https: / / physionet.org / content / mitdb / 1.0.0 / ). [Reference 3] G. Moody and R. Mark, "MIT-BIH Noise Stress Test Database", [Retrieved September 2, 2021], (https: / / physionet.org / content / nstdb / 1.0.0 / ). [Reference 4] JP 2020-036781 A [Explanation of symbols]

[0064] 101...first arithmetic circuit, 102...second arithmetic circuit, 103...third arithmetic circuit, 104...fourth arithmetic circuit, 105...fifth arithmetic circuit, 111...transmission circuit, 112...display device, 121...electrocardiograph.

Claims

1. a first arithmetic circuit that detects R waves from an electrocardiogram waveform measured by an electrocardiograph and determines an interval between adjacent R waves; a second arithmetic circuit for calculating an instantaneous heart rate for each beat represented by an R wave from the interval calculated by the first arithmetic circuit; a third arithmetic circuit for obtaining a smoothed heart rate by smoothing the instantaneous heart rate for each beat calculated by the second arithmetic circuit; a fourth calculation circuit that calculates a signal-to-noise ratio of the electrocardiogram waveform using at least one of the instantaneous heart rate and the smoothed heart rate; a transmission circuit that transmits the signal-to-noise ratio calculated by the fourth calculation circuit to a set destination; Equipped with A sixth arithmetic circuit is further provided for determining the appropriateness of the smoothed heart rate calculated by the third arithmetic circuit, The fourth arithmetic circuit calculates a signal-to-noise ratio of the electrocardiogram waveform based on the result of the determination by the sixth arithmetic circuit. A measurement quality evaluation device characterized by:

2. 2. The measurement quality evaluation device according to claim 1, a fifth arithmetic circuit that determines a period during which the signal-to-noise ratio determined by the fourth arithmetic circuit falls below a certain value within a measurement period during which electrocardiogram information is measured, and displays this period on a display device; or that determines a ratio between this measurement period and a period during which the signal-to-noise ratio determined by the fourth arithmetic circuit falls below a certain value within the measurement period during which electrocardiogram information is measured, and displays this ratio on a display device.

3. 3. The measurement quality evaluation device according to claim 1, a seventh arithmetic circuit that generates non-calculation information when an R wave is not detected for a set period in the first arithmetic circuit, The measurement quality evaluation device is characterized in that the transmission circuit transmits the incomputable information.

4. A first step of detecting R waves from an electrocardiogram waveform measured by an electrocardiograph and determining an interval between adjacent R waves; a second step of calculating an instantaneous heart rate for each beat represented by an R wave from the interval calculated in the first step; a third step of calculating a smoothed heart rate by smoothing the instantaneous heart rate for each beat calculated in the second step; a fourth step of determining a signal-to-noise ratio of the electrocardiogram waveform using the instantaneous heart rate and the smoothed heart rate; a fifth step of transmitting the signal-to-noise ratio obtained in the fourth step to a set destination; Equipped with A sixth step of determining whether the smoothed heart rate calculated in the third step is appropriate, The fourth step includes a step of calculating a signal-to-noise ratio of the electrocardiogram waveform based on the result of the determination in the sixth step. A measurement quality evaluation method comprising:

5. 5. The method for evaluating measurement quality according to claim 4, A seventh step of generating non-calculation information when an R wave is not detected for a set period in the first step, The measurement quality evaluation method, wherein the fifth step includes a step of transmitting the incomputable information.

6. 6. A program for causing a computer to execute the measurement quality evaluation method according to claim 4 or 5.

Images