Blood pressure measurement device and blood pressure measurement method

Abstract

To correctly inform a user of the necessity of calibration.SOLUTION: A blood pressure measurement device according to an aspect comprises: a measurement unit for measuring pulse waves; a blood pressure calculation unit which calculates a blood pressure based on the pulse waves; a signal output unit which receives input of the pulse waves and outputs a signal by a learned model; and a determination unit which determines the propriety of the reproduction of the pulse waves in the signal.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a blood pressure measurement device and a blood pressure measurement method. [Background technology]

[0002] 2. Description of the Related Art Conventionally, blood pressure measuring devices that estimate blood pressure from a pulse wave have been known, and various methods for estimating blood pressure have also been known. For example, Patent Document 1 proposes a technique for calculating blood pressure based on the ratio (v / ΔP) of the wave height (v) of the first peak of the waveform obtained by first-order differentiation of the pulse waveform to the wave height (ΔP) of the rising edge of the first peak in the pulse waveform before differentiation.

[0003] Many of the previously proposed technologies require periodic calibration by comparing the measured value with that of a reference blood pressure monitor due to various factors such as aging, changes in physical constitution, changes in the state of illness, climate change, and deterioration of the device over time. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 4645259 Summary of the Invention [Problem to be solved by the invention]

[0005] However, frequent calibration to maintain the accuracy of blood pressure measurements increases the burden on the user and reduces convenience. On the other hand, infrequent calibration reduces the accuracy of blood pressure measurements. For this reason, it is desirable to accurately notify the user of the need for calibration and perform calibration at appropriate intervals.

[0006] In view of the above circumstances, an object of the present invention is to accurately notify a user of the need for calibration. [Means for solving the problem]

[0007] In order to solve the above problem, a blood pressure measurement device according to one aspect of the present invention includes a measurement unit that measures a pulse wave, a blood pressure calculation unit that calculates blood pressure based on the pulse wave, a signal output unit that receives the pulse wave as input and outputs a signal using a trained model, and a judgment unit that judges whether the pulse wave is reproduced properly in the signal.

[0008] In a blood pressure measurement device according to one aspect of the present invention, the determination unit determines whether the pulse wave has been reproduced properly based on the difference between the calculated blood pressure value based on the pulse wave and the calculated blood pressure value based on the signal. In the blood pressure measurement device according to one aspect of the present invention, the determination unit outputs an instruction to calibrate the blood pressure value when it determines that the reproduction of the pulse wave is poor.

[0009] According to a blood pressure measurement device according to one aspect of the present invention, if it is determined that the reproduction of the pulse wave is poor, the trained model is retrained based on the pulse wave. A blood pressure measurement method according to one aspect of the present invention includes a measurement step of measuring a pulse wave, a blood pressure calculation step of calculating blood pressure based on the pulse wave, a step of inputting the pulse wave and outputting a signal using a trained model, and a step of determining whether the pulse wave is reproduced properly in the signal. [Effects of the Invention]

[0010] According to the present invention, the user can accurately know when calibration is necessary. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 is a schematic diagram illustrating the external configuration of a blood pressure measurement device. [Figure 2] FIG. 2 is a block diagram showing the functional configuration of the blood pressure measurement device. [Figure 3] 1 is a flowchart illustrating an example of a blood pressure measurement method using a blood pressure measurement device. [Figure 4] FIG. 4 is a diagram illustrating an example of feature amounts of a pulse wave signal. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the present invention, and not all of the combinations of features described in the embodiments are necessarily essential to the configuration of the present invention. The configuration of the embodiments may be modified or changed as appropriate depending on the specifications of the device to which the present invention is applied and various conditions (such as usage conditions and usage environment).

[0013] The technical scope of the present invention is defined by the claims and is not limited by the individual embodiments described below. The drawings used in the following description may differ in scale and shape from the actual structure to make each configuration easier to understand. The correspondence between the drawings may also differ in some places to simplify the description. Components shown in previously described drawings may be referenced as appropriate in the description of subsequent drawings.

[0014] <Configuration of blood pressure measurement device> FIG. 1 is a schematic diagram showing the external configuration of a blood pressure measurement device 100. As shown in FIG. The blood pressure measurement device 100 includes a sensor 110 and a main body 120 , and the sensor 110 and the main body 120 are connected by a signal cable 105 .

[0015] As an example, sensor 110 is a fingertip-type sensor that is worn on a user's finger to measure the user's pulse wave and output a pulse wave signal. Sensor 110 has a clamping part 112 that clamps the user's finger to secure sensor 110 to the finger, and a gripping part 111 that the user grips to open and close clamping part 112. Sensor 110 may also be a wristwatch-type sensor that is fastened to the wrist.

[0016] The sensor 110 is, for example, a photoplethysmographic sensor that measures a pulse wave by irradiating a finger with light and receiving light returning from the finger. The blood pressure measurement device 100 may measure a pulse wave using a piezoelectric sensor or an electrocardiogram sensor instead of a photoplethysmographic sensor. Main unit 120 has a built-in computer and the like, and calculates the user's blood pressure based on the pulse wave signal sent from sensor 110 via signal cable 105. Main unit 120 has an operation unit 121 including operation buttons and the like for operating blood pressure measuring device 100, and a display unit 122 that displays blood pressure values ​​and the like.

[0017] FIG. 2 is a block diagram showing the functional configuration of the blood pressure measurement device 100. The blood pressure measurement device 100 includes a sensor 110, which is the above-mentioned photoplethysmographic sensor, and further includes a blood pressure calculation unit 131, a blood pressure display unit 132, an autoencoder 133, a determination unit 134, a result display unit 135, a calibration unit 136, and a relearning unit 137.

[0018] The blood pressure calculation unit 131 analyzes the pulse wave signal sent from the sensor 110 and calculates the blood pressure value. The blood pressure display unit 132 displays the blood pressure value calculated by the blood pressure calculation unit 131 on the display unit 122 of the main body 120.

[0019] The autoencoder 133 compresses and reproduces the pulse wave signal using a trained model and outputs a reproduced signal. The reproduced signal is the signal output from the autoencoder 133. Depending on the characteristics of the input pulse wave, if a pulse wave with similar characteristics is input to the autoencoder 133, a signal with high reproducibility of the pulse wave is output, and if a pulse wave with dissimilar characteristics is input to the autoencoder 133, a signal with low reproducibility of the pulse wave is output, which is why it is called a reproduced signal.

[0020] The blood pressure measurement device 100 may be provided with a support vector machine or the like instead of the autoencoder 133, and may output a reproduction signal using another machine learning model such as a regression model. This is because, like an autoencoder, it is also possible to output a reproduction signal by providing another machine learning model that learns to output the same signal as the input.

[0021] Based on the pulse wave signal and the reproduction signal, the determination unit 134 determines whether the reproduction of the pulse wave signal in the reproduction signal is satisfactory. For example, it calculates a value based on the pulse wave signal and the reproduction signal, determines the value based on a predetermined criterion, and determines whether the reproduction of the pulse wave signal in the reproduction signal is satisfactory. Specific determination content may be, for example, whether the difference value obtained by comparing the pulse wave signal and the reproduction signal exceeds a predetermined threshold, or whether the difference value between the blood pressure value calculated from the pulse wave signal and the blood pressure value calculated from the reproduction signal exceeds a predetermined threshold. If the difference value exceeds the threshold, it is determined to be defective.

[0022] To determine whether the reproduction of the pulse wave signal in the reproduction signal is satisfactory, a value other than the difference value may be calculated based on the pulse wave signal and the reproduction signal, and the quality of the reproduction of the pulse wave signal in the reproduction signal may be determined based on a criterion other than comparing this value with a threshold value. Furthermore, when the determination unit 134 determines that the reproduction of the pulse wave signal in the reproduction signal is poor, it may output an instruction to calibrate the blood pressure calculation unit 131. The output instruction to calibrate may be confirmed by the user to trigger manual calibration, or may be used as an instruction to start automatic calibration.

[0023] The result display unit 135 displays the determination result by the determination unit 134 on the display unit 122 of the main body 120. The determination result may be displayed only when the reproduction is poor, and may not be displayed when the reproduction is good. As the determination result, a pass / fail determination may be displayed directly, or the pass / fail determination may be displayed indirectly by, for example, displaying the necessity of calibration for the blood pressure calculation unit 131.

[0024] The calibration unit 136 receives input of the user's blood pressure value obtained by a standard sphygmomanometer and calibrates the blood pressure calculation value by the blood pressure calculation unit 131. That is, the calibration unit 136 rewrites the calculation formula of the blood pressure calculation unit 131 so that the blood pressure value calculated by the blood pressure calculation unit 131 matches the blood pressure value obtained by the standard sphygmomanometer.

[0025] Re-learning unit 137 uses the user's new pulse wave signal as training data to re-learn the trained model so as to improve the reproducibility of the pulse wave signal in the reproduced signal. The trained model used in the autoencoder 133 is, for example, one that has undergone the following three stages of training. In the first stage of training, pulse wave signals of multiple users are used as training data to train the characteristics of the pulse wave signal. The trained model that has undergone the first stage of training is incorporated into the autoencoder 133 in advance, for example, when the blood pressure measurement device 100 is shipped.

[0026] In the second stage of learning, pulse wave signals measured under multiple measurement conditions for a specific user of the blood pressure measurement device 100 are used as training data to learn the characteristics of the pulse wave signal for the specific user. For example, multiple measurement conditions include measurements under both high and low blood pressure conditions. The second stage of learning is performed by the re-learning unit 137, for example, after purchasing the blood pressure measurement device 100 and before starting to use it.

[0027] In the third stage of learning, the pulse wave signal of a specific user in a normal state is used as training data to learn the characteristics of the pulse wave signal in a normal state. Learning in a normal state differs from learning in the second stage of learning; for example, learning is performed by measuring the blood pressure multiple times when the blood pressure is approximately the same. The third stage of learning is performed by the re-learning unit 137 when the blood pressure measurement device 100 is first used or when the blood pressure calculation unit 131 is calibrated.

[0028] The autoencoder 133 using such a trained model reproduces with high reproducibility pulse wave signals measured under the same conditions as those used at the time of calibration. On the other hand, the reproducibility of the autoencoder 133 decreases for pulse wave signals measured under conditions that have changed since calibration. That is, the pass / fail judgment by the judgment unit 134 corresponds to a judgment of whether there is a change in the measurement conditions, and the user can accurately know the need for calibration or the like.

[0029] In addition, a small model created by distillation from a large-scale model created on an external computer such as the cloud may be used as the trained model. <Blood pressure measurement method>

[0030] FIG. 3 is a flowchart showing an example of a blood pressure measurement method using the blood pressure measurement device 100. In step S101, the pulse wave of the user is measured by sensor 110. Then, in step S102, blood pressure calculation unit 131 calculates a blood pressure value based on the pulse wave signal, and blood pressure display unit 132 displays the blood pressure value.

[0031] In step S103, the pulse wave signal is reproduced by autoencoder 133. Then, in step S104, blood pressure calculation unit 131 calculates the blood pressure value based on the reproduced signal. The blood pressure calculation unit 131 calculates the blood pressure value using, for example, the feature amount of the pulse wave signal.

[0032] FIG. 4 is a diagram illustrating an example of feature quantities of a pulse wave signal. The horizontal axis in FIG. 4 represents time, and the vertical axis represents the signal intensity of the pulse wave signal. Possible features used in calculating blood pressure values ​​include, for example, the pulse wave's rising waveform's wave height ΔP, rise time UT, and maximum slope α. In addition, the time ΔT between the peak of the forward wave and the peak of the reflected wave of the pulse wave can also be used as a feature in calculating blood pressure values.

[0033] The formula for calculating the blood pressure value uses the wave height ΔP, rise time UT, maximum slope α, and time between peaks ΔT, either individually or in combination, as shown in Fig. 4. The formula for calculating the blood pressure value may use feature quantities other than those shown in Fig. 4, and any known formula can be used to calculate the blood pressure value.

[0034] However, the well-known formulas for calculating blood pressure values ​​are not versatile enough to obtain accurate blood pressure values ​​even when the pulse wave measurement conditions change and the shape of the pulse wave changes, so calibration is required due to various factors such as aging, changes in constitution, climate change, and deterioration of the device over time.

[0035] In step S105 of Fig. 3, the determination unit 134 determines whether the difference between the blood pressure value based on the pulse wave signal and the blood pressure value based on the reproduced signal exceeds a predetermined threshold. The threshold may be, for example, a measurement error value acceptable for a sphygmomanometer. If the difference between the blood pressure values ​​does not exceed the predetermined threshold (step S105; No), it is considered that there has been no change in the measurement conditions since the previous calibration, and the blood pressure measurement ends.

[0036] On the other hand, if the difference in blood pressure values ​​exceeds the predetermined threshold (step S105; Yes), it is considered that the measurement conditions have changed since the previous calibration, and the fact that the reproduction is poor is displayed by the result display unit 135. A change in the measurement conditions may be due to improper attachment of the sensor 110, so it is desirable to reattach the sensor 110 and execute the process again from step S101.

[0037] If the difference in blood pressure values ​​exceeds the threshold even after one or more retry attempts, calibration is deemed necessary, and in step S106, the blood pressure values ​​are calibrated by the calibration unit 136, for example, in response to the user's operation of the operation unit 121. Specifically, the calibration of blood pressure values ​​is realized, for example, by adjusting parameter values ​​in a calculation formula for calculating blood pressure values, or by modifying adjustment constants or adjustment coefficients that are added to or multiplied by the calculated blood pressure values.

[0038] When the blood pressure value is calibrated, it is considered that the shape of the user's pulse wave has changed, so the re-learning unit 137 re-learns the learned model of the autoencoder 133 in order to re-learn the characteristics of the changed pulse wave. As described above, the blood pressure measurement device and blood pressure measurement method of the present invention allow the user to accurately know the need for blood pressure calibration. Therefore, the accuracy of blood pressure measurement is maintained with the minimum necessary calibration, which reduces the burden of calibration and is highly convenient.

[0039] In the above description, the blood pressure measurement device 100 including the calibration unit 136 is exemplified, but the blood pressure measurement device of the present invention may be calibrated by being connected to, for example, a personal computer. In the above description, the blood pressure measurement device 100 including the relearning unit 137 is exemplified, but the blood pressure measurement device of the present invention may have a trained model that has been retrained by, for example, a personal computer installed therein. [Explanation of symbols]

[0040] 100 Blood pressure measuring device 110 Sensors 120 body 131 Blood pressure calculation unit 132 Blood pressure display unit 133 Autoencoder 134 Judgment section 135 Result display section 136 Proofreading Department 137 Re-learning Section

Claims

1. a measuring unit that measures a pulse wave; a blood pressure calculation unit that calculates blood pressure based on the pulse wave; a signal output unit that receives the pulse wave and outputs a signal based on a trained model; a determination unit that determines whether the reproduction of the pulse wave in the signal is good or bad; A blood pressure measuring device comprising:

2. The blood pressure measurement device according to claim 1 , wherein the determination unit determines whether the reproduction of the pulse wave is satisfactory based on a difference between a calculated value of blood pressure based on the pulse wave and a calculated value of blood pressure based on the signal.

3. The blood pressure measurement device according to claim 1 , wherein the determining unit outputs an instruction to calibrate the blood pressure value when it determines that the reproduction of the pulse wave is poor.

4. The blood pressure measurement device according to claim 1 , wherein the trained model is retrained based on the pulse wave when the reproduction of the pulse wave is determined to be poor.

5. measuring a pulse wave; calculating a blood pressure based on the pulse wave; A step of receiving the pulse wave and outputting a signal using a trained model; determining whether the pulse wave is reproduced in the signal; A blood pressure measurement method comprising:

Images