JP2020092738A - Blood pressure estimation device, blood pressure estimation system, and blood pressure estimation program - Google Patents

Abstract

To estimate a blood pressure of a person to be measured.SOLUTION: A blood pressure estimation device 100 includes: a feature extraction network for converting input pulse wave data to feature representation; a blood pressure estimation network for converting the feature representation converted by the feature extraction network to a probability vector indicating the probability of a systolic blood pressure value and a diastolic blood pressure value; training step execution means for executing a training step for causing the feature extraction network and the blood pressure estimation network to learn the systolic blood pressure value and the diastolic blood pressure value according to pulse wave data; and estimation step execution means for inputting the pulse wave data on the person to be measured to the feature extraction network and the blood pressure estimation network that have completed learning by the training step, and executing an estimation step for estimating the systolic blood pressure value and the diastolic blood pressure value of the person to be measured.SELECTED DRAWING: Figure 1

Description

The present invention relates to a blood pressure estimation device, a blood pressure estimation system, and a blood pressure estimation program.

The following blood pressure estimation device is known. In this blood pressure estimation device, a method of estimating the blood pressure of the subject from the characteristic amount of the pulse wave using a neural network learned based on the measurement value of the cuff sphygmomanometer and the characteristic amount of the pulse wave has been proposed. (For example, patent document 1).

JP, 2016-16295, A

In recent years, it has become possible for a user to easily acquire the pulse wave of a user with a wearable terminal worn on the wrist or the like. Therefore, a method of estimating the blood pressure from the pulse wave of the user acquired by the wearable terminal or the like using a conventional blood pressure estimation device can be considered. However, in the conventional technique, the feature amount of the pulse wave of the measurement subject is extracted, and the estimated blood pressure is output when the extracted feature amount is input, and is generated by the support vector regression analysis based on the teacher data. The model was used to estimate the blood pressure of the measurement subject. When the blood pressure is estimated based on the pulse wave of the measurement target person, it is difficult to specify the blood pressure value uniformly from the pulse wave, and there is a possibility that it will fluctuate within a certain range depending on the person. Therefore, it is expected that the estimation accuracy will be improved by estimating the blood pressure value according to the pulse wave in consideration of its probability, but no consideration has been made so far regarding this point.

A blood pressure estimation device according to the present invention is a blood pressure estimation device that estimates blood pressure using a neural network, and includes a feature extraction network that transforms input pulse wave data into a feature expression, and a feature that is transformed by the feature extraction network. The blood pressure estimation network that converts the expression into a probability vector that represents the probability of the systolic blood pressure value and the diastolic blood pressure value, and the systolic blood pressure value and the diastolic blood pressure value according to the pulse wave data to the feature extraction network and the blood pressure estimation network. The pulse wave data of the measurement target person is input to the training step executing means for executing the training step for learning, the feature extraction network and the blood pressure estimation network that have been learned by the training step, and the systolic blood pressure of the measurement target person is input. Estimation step executing means for executing an estimation step of estimating the value and the diastolic blood pressure value.
A blood pressure estimation system according to the present invention is a blood pressure estimation system that estimates blood pressure using a neural network, and includes a feature extraction network that transforms input pulse wave data into a feature expression and a feature that is transformed by the feature extraction network. The blood pressure estimation network that converts the expression into a probability vector that represents the probability of the systolic blood pressure value and the diastolic blood pressure value, and the systolic blood pressure value and the diastolic blood pressure value according to the pulse wave data to the feature extraction network and the blood pressure estimation network. Training step executing means for executing a training step for learning, inputting the pulse wave data of the measurement target person to the feature extraction network and the blood pressure estimation network which have been learned by the training step, and the systolic period of the measurement target person. Estimation step executing means for executing an estimation step of estimating a blood pressure value and a diastolic blood pressure value.
A blood pressure estimation program according to the present invention includes a feature extraction network for converting input pulse wave data into a feature expression, and a probability vector representing the probability of the systolic blood pressure value and the diastolic blood pressure value for the feature expression converted by the feature extraction network. A blood pressure estimation program for estimating a blood pressure using a neural network having a blood pressure estimation network for converting into a blood pressure estimation network, wherein a systolic blood pressure value and a diastolic blood pressure value according to pulse wave data are added to a feature extraction network and a blood pressure estimation network. Input the pulse wave data of the measurement target person to the training step execution procedure for executing the training step for learning, and the feature extraction network and the blood pressure estimation network that have been learned by the training step, and then perform the systolic period of the measurement target person. A program for causing a computer to execute an estimation step execution procedure for executing an estimation step of estimating a blood pressure value and a diastolic blood pressure value.

According to the present invention, in the training step, a pre-trained feature extraction network and a blood pressure estimation network are used, and the pulse wave data of the measurement target person is converted into a probability vector representing the probabilities of the systolic blood pressure value and the diastolic blood pressure value. Since the conversion is performed, the blood pressure of the measurement target person can be accurately estimated based on the converted probability vector.

1 is a block diagram showing the configuration of an embodiment of blood pressure estimation device 100. FIG. It is the figure which showed typically the input data Ut in case the number of sensor signals obtained from the wearable terminal is 1 and the length of each signal segment is n. It is the figure which showed typically the input data Ut in case the number of sensor signals obtained from the wearable terminal is 3 and the length of each signal segment is n. It is the figure which showed typically the input data Ut in case the number of sensor signals obtained from the wearable terminal is 4 and the length of each signal segment is n. It is the figure which showed the flow of the training step typically. It is the figure which showed the flow of the estimation step typically. It is a flowchart figure which shows the flow of a training step. It is a flowchart figure which shows the flow of an estimation step.

FIG. 1 is a block diagram showing the configuration of an embodiment of blood pressure estimation device 100 according to the present embodiment. As the blood pressure estimation device 100, for example, a server device, a personal computer, a smartphone, a tablet terminal or the like is used. FIG. 1 is a block diagram showing a configuration of an embodiment in which a personal computer is used as blood pressure estimation device 100 in the present embodiment.

The blood pressure estimation device 100 includes an operation member 101, a control device 102, a storage medium 103, a display device 104, and a connection interface 105.

The operation member 101 includes various devices operated by the operator of the blood pressure estimation device 100, such as a keyboard and a mouse.

The control device 102 includes a CPU, a memory, and other peripheral circuits, and controls the entire blood pressure estimation device 100. The memory forming the control device 102 is a volatile memory such as an SDRAM. This memory is used as a work memory for the CPU to expand the program when the program is executed and as a buffer memory for temporarily recording data. For example, the data read via the connection interface 102 is temporarily recorded in the buffer memory.

The storage medium 103 is a storage medium for recording various data stored in the blood pressure estimation device 100, data of a program to be executed by the control device 102, and the like, and is, for example, an HDD (Hard Disk Drive) or SSD (Solid State). Drive) or the like is used. The program data recorded in the storage medium 103 is provided by being recorded in a recording medium such as a CD-ROM or a DVD-ROM, or provided via a network, and stores the program data acquired by the operator. Installation on the medium 103 allows the controller 102 to execute the program. In the present embodiment, programs and various data used in the processing described below are recorded in the storage medium 103.

The display device 104 is, for example, a liquid crystal monitor, and displays various display data output from the control device 102.

The connection interface 105 is an interface for the blood pressure estimation device 100 to communicate with external devices such as other devices and terminals by communication. For example, the communication interface 105 includes an interface for connecting to a communication line such as a LAN or the Internet, and an interface for short-range wireless communication such as Bluetooth (registered trademark). The blood pressure estimation device 100 can acquire data from a wearable terminal described later via the communication interface 105.

The blood pressure estimation device 100 according to the present embodiment acquires pulse wave data from a device that can measure the pulse wave of the user, and executes processing for estimating the blood pressure of the user based on the acquired pulse wave data. .. In the present embodiment, the device capable of measuring the pulse wave of the user is, for example, a wearable terminal mounted on the wrist of the user and equipped with a sensor for measuring the pulse wave in a portion in close contact with the user's body. Assume

なお、式（１）において、ｗはウェアラブル端末から得られたセンサ信号の数であり、ｎは各信号セグメントの長さを表す。 In the present embodiment, a training step is executed in advance to capture the output data from the wearable terminal and allow the neural network to learn the blood pressure value according to the output signal from the wearable terminal. The output data from the wearable terminal used in the training step is represented by the following equation (1), for example, as a large set Ut of signal segments.

In Expression (1), w is the number of sensor signals obtained from the wearable terminal, and n is the length of each signal segment.

Ut has a configuration that can be input to the neural network, and the segment obtained from the wearable terminal is configured as a Ut segment as shown in FIGS. 2 to 4, for example.

FIG. 2 shows a case in which the wearable terminal has one sensor, that is, the number w of sensor signals obtained from the wearable terminal is one. In FIG. 2, X represents a segment extracted from Sensor1 included in the wearable terminal. If the length of the segment X is n, the Ut segment constitutes a 1×n×1 vector.

FIG. 3 shows a case in which the wearable terminal has three sensors, that is, the number w of sensor signals obtained from the wearable terminal is three. In FIG. 3, X represents a segment extracted from Sensor1 included in the wearable terminal, Y represents a segment extracted from Sensor2 included in the wearable terminal, and Z represents a segment extracted from Sensor3 included in the wearable terminal. There is. When the lengths of the segment X, the segment Y, and the segment Z are n, the Ut segment forms a 1×n×3 vector.

FIG. 4 shows a case in which the wearable terminal includes four sensors, that is, the number w of sensor signals obtained from the wearable terminal is four. In FIG. 4, X represents a segment extracted from Sensor1 included in the wearable terminal, Y represents a segment extracted from Sensor2 included in the wearable terminal, and Z represents a segment extracted from Sensor3 included in the wearable terminal. , G represents a segment extracted from Sensor 4 included in the wearable terminal. When the length of each of the segment X, the segment Y, the segment Z, and the segment G is n, the Ut segment forms a 1×n×4 vector.

In the present embodiment, the control device 102 executes the training step according to the flow shown in FIG. 5, when the training data Ut acquired in advance is input. In FIG. 5, Ut indicates a case where the number of sensor signals obtained from the wearable terminal is 3 and the length of each signal segment is n, as shown in FIG. As shown in FIG. 5, the control device 102 uses the data Ut as input data and inputs the data Ut to the feature extraction network 5a prepared in advance.

The feature extraction network 5a is Convolutional Layers in a known convolutional neural network (Convolutional Neural Network), and uses a feature mapping F(z||Ut) for mapping the input signal Ut to the feature expression z in the neural network. Convert Ut to feature representation z. Here, the characteristic expression is also called a latent expression, and the characteristic expression z is represented by the following expression (2). In the following expression (2), k is the length of the feature expression z.

The feature expression z converted by the feature extraction network 5a is input to the blood pressure estimation network 5b. The blood pressure estimation network 5b is Fully Connected Layers in a known convolutional neural network (Fully Connected Neural Network), and is implemented by using layers that are completely connected using the following equation (3).

The blood pressure estimation network 5b uses the estimation mapping Q((qs,qd)||z for mapping the characteristic expression z to a set of probability vectors qs and qd representing the probabilities of the systolic blood pressure value and the diastolic blood pressure value, respectively. ) Is used to transform the feature expression z into a pair of probability vectors qs and qd for learning. qs is a probability vector showing a systolic blood pressure value, and qd is a probability vector showing a diastolic blood pressure value. For example, the blood pressure estimation network 5b is normalized by the softmax function qi defined by the following equation (4), and converts the output vectors qs and qd into probability vectors whose sum is 1.

The feature extraction network 5a and the blood pressure estimation network 5b repeat training while adjusting network weights until the loss function L(ps, pd, qs, qd) becomes minimum. Note that ps and pd are probability vectors indicating the actual systolic blood pressure value and diastolic blood pressure value for the training data Ut. That is, for the person who acquired the training data Ut, the actual systolic blood pressure value and the diastolic blood pressure value are acquired using a known blood pressure monitor, and the probability vectors indicating the values are set as ps and pd. You can use it. Thereby, the feature extraction network 5a and the blood pressure estimation network 5b can be learned until the difference between the blood pressure value estimated using the feature extraction network 5a and the blood pressure estimation network 5b and the actually measured blood pressure value becomes the minimum. The accuracy of blood pressure estimation in the estimation step described below can be improved.

Also, the weights of the network can be adjusted using any learning method such as gradient descent, stochastic gradient descent, or simulated annealing. For example, in the present embodiment, the loss function L(ps, pd, qs, qd) is defined by the following equation (5).

When the training steps described above are completed, the estimation step for estimating the blood pressure of the user can be executed based on the signal input from the wearable terminal.

In the estimation step, the control device 102 inputs the signal U′t represented by the following expression (6), which is input from the wearable terminal, into the network including the feature extraction network 5a and the blood pressure estimation network 5b described above.

As a result, a new segment feature representation z′ and probability vectors qs′ and qd′ for the input signal signal U′t can be obtained. Note that FIG. 6 is a diagram schematically showing the flow from the input of the input signal signal U′t to the output of the probability vectors qs′ and qd′.

Each entry of the output vector output by the blood pressure estimation network 5b represents the probability that the input segment corresponds to the systolic blood pressure value and the diastolic blood pressure value, respectively. Therefore, in the present embodiment, the values having the highest probabilities are estimated as the systolic blood pressure and the diastolic blood pressure of the measurement target person using the following equations (7) and (8).

For example, when a blood pressure value of (to 49 mmHG, 50 mmHG to 89 mmHG, 90 mmHG) is set for the probability vector and qs=(0.3, 0.8, 0.0), the systolic blood pressure value is set. Is estimated to be 50 mmHG to 89 mmHG for which the highest probability is calculated. When qd=(0.0, 0.44, 0.56), the diastolic blood pressure value is estimated to be 90 mmHG or higher for which the highest probability is calculated. In equations (7) and (8), the probability vectors qs and qd are probability vectors that represent the probabilities of estimation that the segment has the systolic blood pressure value sbp′ and the diastolic blood pressure value dbp′, respectively.

なお、式（９）において、確率ベクトルｐｓおよびｐｄは、セグメントが収縮期血圧値ｓｂｐおよび拡張期血圧値ｄｂｐをそれぞれ有する既知の確率を表す確率ベクトルである。 In the present embodiment, the following equation (9) is used to convert the blood pressure value (mmHG) into a vector.

In equation (9), the probability vectors ps and pd are probability vectors that represent known probabilities that the segment has the systolic blood pressure value sbp and the diastolic blood pressure value dbp, respectively.

With the above processing, the blood pressure of the user wearing the wearable terminal can be accurately estimated based on the signal U't input from the wearable terminal.

FIG. 7 is a flowchart showing the flow of training steps in this embodiment. The processing illustrated in FIG. 7 is executed by the control device 102 as a program to be started when the training data Ut acquired in advance is input.

In step S10, the control device 102 inputs the input data Ut to the feature extraction network 5a. As a result, the input signal Ut is converted into the feature representation z by the feature extraction network 5a using the above-described feature mapping F(z||Ut). Then, it progresses to step S20.

In step S20, the blood pressure estimation network 5b converts the feature expression z into a pair of probability vectors qs and qd using the estimated mapping Q((qs,qd)||z). Then, it progresses to step S30.

In step S30, the control device 102 calculates the loss function L (ps, pd, qs, qd) described above. Then, it progresses to step S40.

In step S40, the control device 102 determines whether or not the above-mentioned loss function L has become minimum. If an affirmative decision is made in step S40, the weights at that time are adopted as the feature extraction network 5a and the blood pressure estimation network 5b, and the processing ends. On the other hand, if a negative decision is made in step S40, the operation proceeds to step S50.

In step S50, the control device 102 adjusts the weights of the feature extraction network 5a and the blood pressure estimation network 5b, as described above, and returns to step S10.

FIG. 8 is a flowchart showing the flow of the estimation step in this embodiment. The process shown in FIG. 8 is executed by the control device 102 as a program to be started when the data Ut′ of the measurement target person is input.

In step S110, the control device 102 inputs the input data Ut' to the feature extraction network 5a. Thereby, the feature extraction network 5a transforms each signal segment of the input signal Ut' into the feature representation z using the feature mapping F(z||Ut) described above. Then, it progresses to step S120.

In step S120, the blood pressure estimation network 5b converts the feature expression z into a pair of probability vectors qs' and qd' using the estimated mapping Q((qs,qd)||z). Then, it progresses to step S130.

In step S130, the control device 102 calculates the systolic blood pressure value and the diastolic blood pressure value of the measurement target person from the probability vectors qs′ and qd as described above. Then, it progresses to step S140.

In step S140, the control device 102 outputs and displays the calculated systolic blood pressure value and diastolic blood pressure value of the measurement subject on the display device 104. Then, the process ends.

According to the embodiment described above, the following operational effects can be obtained.
(1) A feature extraction network that transforms input pulse wave data into a feature expression, and blood pressure estimation that transforms the feature expression converted by the feature extraction network into a probability vector that represents the probability of a systolic blood pressure value and a diastolic blood pressure value. The control device 102 includes a network, and the control device 102 executes a training step for causing the feature extraction network and the blood pressure estimation network to learn the systolic blood pressure value and the diastolic blood pressure value according to the pulse wave data, and the learning step performs learning. The pulse wave data of the measurement subject is input to the completed feature extraction network and blood pressure estimation network, and the estimation step of estimating the systolic blood pressure value and the diastolic blood pressure value of the measurement subject is executed. This makes it possible to accurately measure the blood pressure of the measurement target person by using the feature extraction network and the blood pressure estimation network that have been trained in advance.

(2) The control device 102 obtains a probability vector representing the probabilities of the systolic blood pressure value and the diastolic blood pressure value of the measurement target person, and sets the value having the highest probability to the systolic blood pressure value and the diastolic phase of the measurement target person. It was estimated as a blood pressure value. Thereby, the measurement accuracy of blood pressure can be improved.

(3) The control device 102 displays the estimated systolic blood pressure value and diastolic blood pressure value of the measurement subject on the display device 104. Thereby, an operator such as a doctor can confirm the blood pressure value of the measurement target person on the display device 104.

(4) In the training step, the feature extraction network 5a and the blood pressure estimation network 5b repeat the training while adjusting the network weights until the loss function L(ps, pd, qs, qd) becomes minimum. This can improve the accuracy of blood pressure estimation based on the pulse wave data.

-Modification-
The blood pressure estimation system of the above-described embodiment can be modified as follows.
(1) In the above-described embodiment, the blood pressure estimation device 100 is a personal computer, and the control device 102 executes the above-described processing. However, the terminal operated by the operator is allowed to connect to the blood pressure estimation apparatus 100 via a communication line such as the Internet, and the pulse wave data from the wearable terminal is input by being transmitted to the blood pressure estimation apparatus 100 from the operation terminal. May be done. In that case, the blood pressure estimation result may be transmitted to the operation terminal and displayed on the operation terminal. With this, by connecting the operation terminal and the blood pressure estimating apparatus 100 via the communication line, it is possible to construct a client server type or cloud type blood pressure estimating system, while using the blood pressure estimating apparatus 100 alone. It can also be used as a stand-alone device.

(2) In the above-described embodiment, the example in which the control device 102 of the blood pressure estimation device 100 executes the processing for the training step has been described. However, the processing for the training step may be executed by another device, and the data obtained by the training step may be recorded in the storage medium 103. In this case, the processing for the training step in the blood pressure estimation device 100 becomes unnecessary.

(3) In the above-described embodiment, the feature extraction network 5a and the blood pressure estimation network 5b repeat training while adjusting network weights until the loss function L(ps, pd, qs, qd) is minimized. explained. However, the feature extraction network 5a and the blood pressure estimation network 5b may repeat training while adjusting the weights of the networks until the loss function L(ps, pd, qs, qd) becomes equal to or less than a preset threshold value. ..

The present invention is not limited to the configurations of the above-described embodiments as long as the characteristic functions of the present invention are not impaired. Further, a configuration in which the above-described embodiment and a plurality of modified examples are combined may be adopted.

100 blood pressure estimation device 101 operation member 102 control device 103 storage medium 104 display device 105 connection interface

A blood pressure estimation device according to the present invention is a blood pressure estimation device that estimates blood pressure using a neural network, and includes a feature extraction network that transforms input pulse wave data into a feature expression, and a feature that is transformed by the feature extraction network. The blood pressure estimation network that converts the expression into a probability vector that represents the probabilities of the systolic blood pressure value and the diastolic blood pressure value, and the systolic blood pressure value and the diastolic blood pressure value according to the pulse wave data in the feature extraction network and the blood pressure estimation network. The pulse wave data of the measurement target person is input to the training step executing means for executing the training step for learning, the feature extraction network and the blood pressure estimation network that have been learned by the training step, and the systolic blood pressure of the measurement target person is input. And an estimation step executing means for executing an estimation step for estimating a value and a diastolic blood pressure value , the estimation step executing means obtaining a probability vector representing probabilities of the systolic blood pressure value and the diastolic blood pressure value of the measurement subject. , The value having the highest probability is estimated as the systolic blood pressure value and the diastolic blood pressure value of the measurement subject .
A blood pressure estimation system according to the present invention is a blood pressure estimation system that estimates blood pressure using a neural network, and includes a feature extraction network that transforms input pulse wave data into a feature expression and a feature that is transformed by the feature extraction network. The blood pressure estimation network that converts the expression into a probability vector that represents the probabilities of the systolic blood pressure value and the diastolic blood pressure value, and the systolic blood pressure value and the diastolic blood pressure value according to the pulse wave data in the feature extraction network and the blood pressure estimation network. Training step executing means for executing a training step for learning, inputting the pulse wave data of the measurement target person to the feature extraction network and the blood pressure estimation network which have been learned by the training step, and the systolic period of the measurement target person. An estimation step executing means for executing an estimation step for estimating a blood pressure value and a diastolic blood pressure value, and the estimation step executing means obtains a probability vector representing probabilities of the systolic blood pressure value and the diastolic blood pressure value of the measurement subject. Then, the value having the highest probability is estimated as the systolic blood pressure value and the diastolic blood pressure value of the measurement subject .
A blood pressure estimation program according to the present invention includes a feature extraction network for converting input pulse wave data into a feature expression, and a probability vector representing the probability of the systolic blood pressure value and the diastolic blood pressure value for the feature expression converted by the feature extraction network. A blood pressure estimation program for estimating a blood pressure using a neural network having a blood pressure estimation network for converting into a blood pressure estimation network, wherein a systolic blood pressure value and a diastolic blood pressure value according to pulse wave data are added to a feature extraction network and a blood pressure estimation network. Input the pulse wave data of the measurement target person to the training step execution procedure for executing the training step for learning, and the feature extraction network and the blood pressure estimation network that have been learned by the training step, and then perform the systolic period of the measurement target person. An estimation step execution procedure for executing an estimation step of estimating a blood pressure value and a diastolic blood pressure value is executed by a computer, and the estimation step execution procedure is a probability representing a probability of a systolic blood pressure value and a diastolic blood pressure value of a measurement target person. It is characterized in that the vector is obtained and the value having the highest probability is estimated as the systolic blood pressure value and the diastolic blood pressure value of the measurement subject .

Claims (12)

A blood pressure estimation device for estimating blood pressure using a neural network,
A feature extraction network that converts the input pulse wave data into a feature expression,
A blood pressure estimation network that converts the feature representation converted by the feature extraction network into a probability vector that represents the probability of a systolic blood pressure value and a diastolic blood pressure value;
Training step executing means for executing a training step for causing the feature extraction network and the blood pressure estimation network to learn a systolic blood pressure value and a diastolic blood pressure value according to pulse wave data,
An estimation step of estimating the systolic blood pressure value and the diastolic blood pressure value of the measurement target person by inputting the pulse wave data of the measurement target person into the feature extraction network and the blood pressure estimation network that have been learned by the training step. A blood pressure estimation device, comprising: an estimation step executing means for executing. The blood pressure estimation device according to claim 1,
The estimating step executing means obtains a probability vector representing the probabilities of the systolic blood pressure value and the diastolic blood pressure value of the measurement target person, and the value having the highest probability is used as the systolic blood pressure value and diastolic value of the measurement target person. A blood pressure estimating device, which estimates as a term blood pressure value. The blood pressure estimation device according to claim 1,
The blood pressure estimation device further comprising display means for displaying the systolic blood pressure value and the diastolic blood pressure value of the measurement subject estimated by the estimation step executing means on a display device. The blood pressure estimation device according to any one of claims 1 to 3,
The blood pressure estimation device, wherein the training step execution means repeats training while adjusting the weights of the feature extraction network and the blood pressure estimation network until the loss function becomes minimum. A blood pressure estimation system for estimating blood pressure using a neural network,
A feature extraction network that converts the input pulse wave data into a feature expression,
A blood pressure estimation network that converts the feature representation converted by the feature extraction network into a probability vector that represents the probability of a systolic blood pressure value and a diastolic blood pressure value;
Training step executing means for executing a training step for causing the feature extraction network and the blood pressure estimation network to learn a systolic blood pressure value and a diastolic blood pressure value according to pulse wave data,
An estimation step of estimating the systolic blood pressure value and the diastolic blood pressure value of the measurement target person by inputting the pulse wave data of the measurement target person into the feature extraction network and the blood pressure estimation network that have been learned by the training step. A blood pressure estimation system comprising: an estimation step executing means for executing. The blood pressure estimation system according to claim 5,
The estimating step executing means obtains a probability vector representing the probabilities of the systolic blood pressure value and the diastolic blood pressure value of the measurement target person, and the value having the highest probability is used as the systolic blood pressure value and diastolic value of the measurement target person. A blood pressure estimation system characterized by estimating as a term blood pressure value. The blood pressure estimation system according to claim 5 or 6,
The blood pressure estimation system further comprising display means for displaying a systolic blood pressure value and a diastolic blood pressure value of the measurement subject estimated by the estimation step execution means on a display device. The blood pressure estimation system according to any one of claims 5 to 7,
The blood pressure estimation system, wherein the training step execution means repeats training while adjusting the weights of the feature extraction network and the blood pressure estimation network until the loss function becomes minimum. A feature extraction network for converting input pulse wave data into a feature expression, and a blood pressure estimation network for converting the feature expression converted by the feature extraction network into a probability vector representing probabilities of systolic blood pressure value and diastolic blood pressure value. A blood pressure estimation program for estimating blood pressure using a neural network having
In the feature extraction network and the blood pressure estimation network, a training step execution procedure for executing a training step for learning a systolic blood pressure value and a diastolic blood pressure value according to pulse wave data,
An estimation step of estimating the systolic blood pressure value and the diastolic blood pressure value of the measurement target person by inputting the pulse wave data of the measurement target person into the feature extraction network and the blood pressure estimation network that have been learned by the training step. And a blood pressure estimation program for causing a computer to execute an estimation step execution procedure. In the blood pressure estimation program according to claim 9,
The estimation step execution procedure obtains a probability vector representing the probabilities of the systolic blood pressure value and the diastolic blood pressure value of the measurement target person, and sets the value having the highest probability to the systolic blood pressure value and diastolic value of the measurement target person. A blood pressure estimation program which estimates as a term blood pressure value. The blood pressure estimation program according to claim 9,
The blood pressure estimation program further comprising a display procedure for displaying the systolic blood pressure value and the diastolic blood pressure value of the measurement subject estimated in the estimation step execution procedure on a display device. The blood pressure estimation program according to any one of claims 9 to 11,
The blood pressure estimation program, wherein the training step execution procedure repeats training while adjusting the weights of the feature extraction network and the blood pressure estimation network until the loss function becomes minimum.

Images