JP2022133919A - Biological information arithmetic system - Google Patents

Abstract

To provide a biological information arithmetic system that can generate an evaluation result related to a state of a user's health with high accuracy.SOLUTION: A biological information arithmetic system generates an evaluation result related to a state of a user's health, and the biological information arithmetic system comprises: acquisition means that acquires evaluation data based on the user's pulse wave; a database that, with a pair of input data based on a pulse wave for learning acquired in advance and reference data including a blood sugar level linked to the input data as data for learning, stores classification information generated by using the plurality of pieces of data for learning; generation means that refers to the database and generates a blood sugar level evaluation result including the blood sugar level for the evaluation data; and evaluation means that, on the basis of the evaluation data and the blood sugar level evaluation result, generates the evaluation result.SELECTED DRAWING: Figure 1

Description

The present invention relates to a biometric information computing system that generates evaluation results regarding the health condition of a user.

Conventionally, as a method for generating a user's blood sugar level, for example, a method as disclosed in Patent Document 1 has been proposed.

In Patent Document 1, the subject's accelerated pulse wave is measured, and from the waveform information of the measured accelerated pulse wave, spectroscopic analysis is performed based on the correlation between the blood sugar level measured by the invasive measurement method and the simultaneously measured accelerated pulse wave. Disclosed is a biological information estimation device and method for extracting blood sugar level information of a subject by a non-invasive method without using

JP 2019-141583 A

Here, in the pulse wave waveform signal, there may be variations in the numerical value of the blood sugar level that can be generated from the pulse wave waveform signal depending on the user's attributes such as gender and age. For example, when comparing the case of processing the measured pulse wave with a male user in his 20s and the case of processing the measured pulse wave with a female user in his 50s, depending on the processing method, blood sugar that can be generated There is a concern that the accuracy of the values may vary. That is, in order to obtain sufficiently accurate evaluation results regarding the user's health condition using the blood glucose level generated from the pulse wave waveform signal, it is necessary to perform processing that matches the characteristics of the user.

On the other hand, Patent Document 1 does not disclose that the blood sugar level generated from the waveform signal of the pulse wave is processed according to the characteristics of the user to generate an evaluation result regarding the health condition of the user. For this reason, in the conventional technology such as Patent Document 1, there is a concern that the evaluation result regarding the user's health condition cannot be generated with high accuracy because the processing that matches the characteristics of the user is not assumed.

Accordingly, the present invention has been devised in view of the above-described problems, and its object is to provide a biological information calculation system capable of generating highly accurate evaluation results regarding the user's health condition. to do.

A biological information computing system according to a first aspect of the present invention is a biological information computing system that generates an evaluation result regarding a user's health condition, comprising an acquisition means for acquiring evaluation data based on the user's pulse wave, and a pre-acquired learning A database storing classification information generated using a plurality of pieces of the learning data, with a pair of input data based on a pulse wave and reference data including a blood glucose level linked to the input data as learning data; generating means for referring to the database and generating a blood sugar level evaluation result including the blood sugar level for the evaluation data; and evaluation means for generating the evaluation result based on the evaluation data and the blood sugar level evaluation result. characterized by

A biological information computing system according to a second invention is characterized in that, in the first invention, the evaluation means detects a tendency of the user's blood glucose level from time-series changes in a plurality of the blood glucose level evaluation results generated from the evaluation data different from each other. and generating the evaluation result from the blood glucose trend and the evaluation data.

A biological information computing system according to a third invention is the biological information computing system according to the first invention or the second invention, wherein the evaluation means selects a function for generating the evaluation result from the blood glucose level evaluation result based on the characteristics of the evaluation data. characterized by

A biological information computing system according to a fourth invention is the biological information computing system according to any one of the first to third inventions, wherein the evaluation means acquires additional information indicating characteristics of the user, and obtains the evaluation data, the blood glucose level evaluation result, and It is characterized by including generating the evaluation result based on the additional information.

A biological information computing system according to a fifth aspect of the present invention is, in any one of the first to third aspects of the invention, wherein the input data includes learning additional information indicating characteristics associated with the learning pulse wave, and the generating means is characterized by acquiring additional information indicating characteristics of the user, referring to the database, and generating the blood glucose level evaluation result for the evaluation data and the additional information.

A biological information computing system according to a sixth aspect of the invention is the biological information computing system according to any one of the first to third aspects of the invention, wherein the generating means acquires additional information indicating characteristics of the user; selecting first classification information based on the first classification information, referring to the first classification information, and generating the blood glucose level evaluation result for the evaluation data.

A biological information computing system according to a seventh aspect of the invention is the fourth aspect of the invention, wherein the acquiring means acquires additional data indicating characteristics different from the evaluation data based on the pulse wave, and the generating means comprises: and generating the additional information based on the additional data.

A biological information computing system according to an eighth invention is the first invention, wherein the acquisition means acquires preliminary evaluation data different from the evaluation data based on the pulse wave, and the classification information is different A plurality of attribute-based classification information calculated using the learning data is included, and the generating means refers to the preliminary evaluation data and selects means for selecting first classification information from among the plurality of attribute-based classification information. and attribute-by-attribute generation means for generating the blood glucose level evaluation result for the evaluation data by referring to the first classification information.

A biological information calculation system according to a ninth aspect of the present invention is characterized in that, in the eighth aspect of the invention, the acquisition means acquires data corresponding to a velocity pulse wave based on the pulse wave as the evaluation data, and an acceleration pulse wave based on the pulse wave. is obtained as the preliminary evaluation data.

According to the first to ninth inventions, the generating means refers to the database and generates the blood sugar level evaluation result including the blood sugar level for the evaluation data. Also, the evaluation means generates an evaluation result based on the evaluation data and the blood glucose level evaluation result. Therefore, the evaluation result can be generated from the evaluation data obtained based on the pulse wave characteristics including the user's characteristics and the blood glucose level evaluation result. As a result, it is possible to highly accurately generate an evaluation result regarding the user's health condition using a processing method suited to the characteristics of the user.

Further, according to the first to ninth inventions, the generating means refers to the database and generates blood glucose level evaluation results for the evaluation data. The database also stores classification information generated using a plurality of pieces of learning data. Therefore, when generating an evaluation result, it is possible to include a quantitative blood sugar level based on data that has been proven in the past. This makes it possible to improve the accuracy when generating evaluation results.

In particular, according to the second invention, the evaluation means acquires a blood sugar trend related to the tendency of the user's blood sugar level from time-series changes in a plurality of blood sugar level evaluation results generated from different evaluation data, and obtains the blood sugar trend and the evaluation. Generate evaluation results from data. Therefore, it is possible to easily grasp characteristics of blood sugar levels, such as blood sugar level spikes, which are different for each user. This makes it possible to generate user evaluation results with high accuracy.

Particularly, according to the third invention, the evaluation means selects a function for generating the evaluation result from the blood glucose level evaluation result based on the characteristics of the evaluation data. Therefore, an appropriate function can be selected according to the characteristics of the evaluation data that differ from user to user. This makes it possible to generate user evaluation results with high accuracy.

In particular, according to the fourth invention, the evaluation means acquires additional information indicating user characteristics, and generates an evaluation result based on the evaluation data, the blood sugar level evaluation result, and the additional information. Therefore, by using a plurality of types of information, it is possible to realize comprehensive evaluation based on multifaceted viewpoints. This makes it possible to generate user evaluation results with higher accuracy.

In particular, according to the fifth invention, the generation means acquires additional information indicating user characteristics, refers to the database, and generates the blood glucose level evaluation result for the evaluation data and the additional information. For this reason, it is possible to use a parameter correlated with the user's blood sugar level evaluation result and realize a comprehensive evaluation based on multiple viewpoints. This makes it possible to generate user evaluation results with higher accuracy.

In particular, according to the sixth aspect, the generation means acquires additional information indicating the biometric information of the user, selects the first classification information from among the classification information based on the additional information, refers to the first classification information, A blood glucose level evaluation result is generated for the evaluation data. Therefore, appropriate classification information can be selected according to the characteristics of the user. This makes it possible to generate user evaluation results with higher accuracy.

In particular, according to the seventh invention, the acquiring means acquires additional data indicating characteristics different from the evaluation data based on the pulse wave, and the generating means generates additional information based on the additional data. including. Therefore, it is possible to use additional information generated from a viewpoint different from the blood sugar level, and it is possible to generate multifaceted evaluation results in consideration of user's characteristics. This makes it possible to generate user evaluation results with higher accuracy.

In particular, according to the eighth invention, the generation means includes selection means for referring to the preliminary evaluation data and selecting the first classification information, and attribute and separately generated information. Therefore, it is possible to select the attribute classification information that is most suitable for the characteristics of the pulse wave, and then generate the blood glucose level evaluation result for the evaluation data. This makes it possible to further improve the evaluation accuracy.

In particular, according to the ninth aspect, the acquisition means acquires data corresponding to the velocity pulse wave based on the pulse wave as the evaluation data, and acquires data corresponding to the acceleration pulse wave based on the pulse wave as the preliminary evaluation data. do. Therefore, the attribute classification information can be selected using the acceleration pulse wave, which makes it easier to classify the characteristics of the pulse wave than the velocity pulse wave. In addition, the blood sugar level evaluation result can be generated using the velocity pulse wave, which makes it easier to calculate the blood sugar level than the accelerated pulse wave. This makes it possible to further improve the evaluation accuracy.

FIG. 1 is a schematic diagram showing an example of a biological information computing system according to the first embodiment. FIGS. 2(a) and 2(b) are schematic diagrams showing an example of the operation of the biological information computing system according to the first embodiment. FIG. 3(a) is a schematic diagram showing an example of classification information, and FIGS. 3(b) and 3(c) are schematic diagrams showing an example of processing for sensor data. FIG. 4(a) is a schematic diagram showing an example of the configuration of the biometric information computing device, and FIG. 4(b) is a schematic diagram showing an example of the functions of the biometric information computing device. FIGS. 5(a) and 5(b) are schematic diagrams showing an example of a sensor. FIG. 6 is a flow chart showing an example of the operation of the biological information computing system according to the first embodiment. FIG. 7 is a schematic diagram showing a classification example of data corresponding to an accelerated pulse wave. FIG. 8 is a schematic diagram showing an example of classification of data corresponding to velocity pulse waves. FIG. 9 is a schematic diagram showing an example of the operation of the biological information computing system according to the second embodiment. FIG. 10 is a schematic diagram showing a blood sugar trend. FIG. 11 is a schematic diagram showing an example of the operation of the biological information computing system according to the third embodiment. FIG. 12 is a schematic diagram showing an example of the operation of the biological information computing system according to the fourth embodiment. FIG. 13 is a schematic diagram showing a modification of the operation of the biological information computing system according to the fifth embodiment. FIGS. 14A and 14B are schematic diagrams showing an example of processing for calculating additional information for sensor data. FIG. 15 is a schematic diagram showing an example of the operation of the biological information computing system according to the sixth embodiment. FIG. 16 is a schematic diagram showing an example of the operation of the biological information computing system according to the seventh embodiment. FIG. 17 is a schematic diagram showing an example of the operation of the biological information computing system according to the eighth embodiment.

An example of a biometric information computing system according to an embodiment of the present invention will be described below with reference to the drawings.

(First embodiment: biological information computing system 100)
FIG. 1 is a schematic diagram showing an example of a biological information computing system 100 according to the first embodiment.

The biometric information computing system 100 is used to generate an evaluation result regarding the user's health condition. The user's health condition evaluation result indicates the result of the user's health checkup, an index indicating the degree of health, insurance premiums according to the user's health condition, and the like.

For example, as shown in FIG. 1, the biometric information computing system 100 includes a biometric information computing device 1, and may include at least one of the sensor 5 and the server 4, for example. The biological information computing device 1 is connected to the sensor 5 and the server 4 via the communication network 3, for example.

The biological information calculation system 100 generates a blood sugar level evaluation result including a blood sugar level from the user's pulse wave-based evaluation data, and generates an evaluation result from the evaluation data and the blood sugar level evaluation result.

In the biometric information computing system 100, for example, as shown in FIG. 2A, the biometric information computing device 1 acquires sensor data generated by the sensor 5 or the like. After that, the biological information computing device 1 performs preprocessing such as filtering on the acquired sensor data, and acquires evaluation data.

The biological information computing device 1 generates a blood sugar level evaluation result for the evaluation data, and generates an evaluation result for the evaluation data and the blood sugar level evaluation result. Therefore, the evaluation result can be generated from the evaluation data obtained based on the pulse wave characteristics and the blood glucose level evaluation result. This makes it possible to generate user evaluation results with high accuracy.

Here, the biological information computing device 1 refers to the database when generating the blood glucose level evaluation result for the evaluation data. The database stores classification information generated using a plurality of pieces of learning data.

Classification information, for example, as shown in FIG. is generated using the training data of Therefore, when the blood sugar level evaluation result is generated, it is possible to include the quantitative blood sugar level based on the input data and the output data that have been proven in the past. This makes it possible to improve the accuracy when generating evaluation results.

The biological information computing device 1 outputs the generated evaluation result to a display or the like, for example.

Incidentally, in the biological information computing system 100, evaluation data may be obtained from the sensor 5 or the like, as shown in FIG. 2B, for example. In this case, preprocessing for obtaining evaluation data from sensor data is performed by the sensor 5 or the like.

<Sensor data>
The sensor data includes data indicating characteristics of the user's pulse wave, and may include, for example, data (noise) indicating characteristics other than the pulse wave. The sensor data is data that indicates the amplitude with respect to the measurement time, and by performing filter processing according to the application and the generation conditions of the sensor data, data corresponding to the acceleration pulse wave, velocity pulse wave, etc. can be obtained from the sensor data. be able to.

The sensor data can be generated by known sensors such as strain sensors, gyro sensors, photoplethysmographic (PPG) sensors, pressure sensors, and the like. The sensor data may be, for example, an analog signal as well as a digital signal. The measurement time when generating sensor data is, for example, the measurement time for 1 to 20 cycles of the pulse wave, and can be set arbitrarily according to the conditions such as the sensor data processing method and data communication method. can be done.

<Evaluation data>
The evaluation data indicates blood glucose level evaluation results and data for generating the evaluation results. The evaluation data indicates, for example, data corresponding to the acceleration pulse wave based on the user's pulse wave, and indicates the amplitude for a specific cycle (for example, one cycle).

The evaluation data is obtained by performing processing (preprocessing) on the sensor data using the biological information computing device 1 or the like. For example, as shown in FIGS. 3B and 3C, evaluation data can be obtained by performing a plurality of processes on sensor data. Details of each process will be described later.

<Database>
The database is mainly used when generating evaluation results for evaluation data. In addition to storing one or more pieces of classification information, the database may also store, for example, a plurality of pieces of learning data used to generate the classification information.

The classification information is, for example, a function indicating the correlation between past evaluation data (input data) obtained in advance and reference data including blood sugar levels. The classification information indicates a calibration model generated based on the results of analysis by regression analysis or the like, for example, using input data as explanatory variables and reference data as objective variables. For example, the classification information can periodically update the calibration model, or may be generated according to attribute information such as the user's sex, age, and exercise content.

Regression analysis methods used to generate classification information include, for example, PLS (Partial Least Squares) regression analysis and SIMCA (Soft Independent Modeling of Class Analogy) method that obtains a principal component model by performing principal component analysis for each class. Regression analysis or the like can be used.

The classification information may include a trained model generated by machine learning using a plurality of learning data, for example. A trained model indicates, for example, a neural network model such as a CNN (Convolutional Neural Network), or an SVM (Support Vector Machine). As machine learning, for example, deep learning can be used.

The input data uses the same type of data as the evaluation data, and indicates, for example, past evaluation data with a clear corresponding blood sugar level. For example, the subject wears the sensor 5 or the like to generate sensor data (sensor data for learning) indicating characteristics of the pulse wave for learning. Then, input data can be acquired by performing processing on the learning sensor data. The input data may be acquired from the user of the biometric information computing system 100, or may be acquired from a user other than the user, for example. In other words, the subject described above may be a user of the biometric information computing system 100, or may be a subject other than the user, and may be a specified or unspecified number of subjects.

It is preferable that the input data be obtained with the same content as the type of the sensor 5 or the like used when obtaining the evaluation data, the generation conditions of the sensor data, and the processing conditions for the sensor data. For example, by unifying the above three contents, it is possible to dramatically improve the accuracy when generating evaluation results.

The reference data includes the subject's blood sugar level measured using a measuring device or the like. For example, when the sensor 5 or the like is attached to a subject to generate learning sensor data, by measuring the blood sugar level of the subject, reference data linked to the input data can be obtained. In this case, the timing of measuring the blood sugar level is preferably simultaneous with the timing of generating the learning sensor data, but it may be, for example, about 1 to 10 minutes before or after.

Reference data is measured using a known measuring device. For example, when measuring a blood sugar level, a known device such as CareFact Link (manufactured by Nipro Corporation) is used as a measuring device.

<Blood sugar level evaluation result>
The blood sugar level evaluation result is generated as the same type of data as the reference data, and includes the blood sugar level. The blood glucose level evaluation result refers to the classification information and is generated as data identical or similar to the reference data.

The biological information computing system 100 acquires, for example, a plurality of evaluation data along an arbitrary time series, and generates a plurality of blood glucose level evaluation results for each evaluation data. Further, in the biological information computing system 100, a plurality of evaluation data may be acquired at arbitrary timing such as when the amount of exercise changes, and a plurality of blood glucose level evaluation results may be generated for each evaluation data.

<Evaluation results>
The evaluation result indicates information about the health condition of the user. Evaluation results include, for example, a plurality of blood sugar level evaluation results and evaluation data, a differential value between a pair of evaluation data and blood sugar level evaluation results, and a known threshold derived based on a plurality of evaluation data and blood sugar level evaluation results. It's okay. By outputting the evaluation result, the health condition of the user can be grasped.

<Biological information computing device 1>
The biological information computing device 1 is an electronic device such as a personal computer (PC), a mobile phone, a smart phone, a tablet terminal, a wearable terminal, etc., for example, an electronic device capable of communicating via the communication network 3 based on a user's operation. Indicates equipment. In addition, the biological information computing device 1 may incorporate the sensor 5 . An example in which a PC is used as the biological information computing device 1 will be described below.

FIG. 4A is a schematic diagram showing an example of the configuration of the biometric information computing device 1, and FIG. 4B is a schematic diagram showing an example of the functions of the biometric information computing device 1. As shown in FIG.

For example, as shown in FIG. 4A, the biometric information computing device 1 includes a housing 10, a CPU (Central Processing Unit) 101, a ROM (Read Only Memory) 102, a RAM (Random Access Memory) 103, A storage unit 104 and I/Fs 105 to 107 are provided. Each configuration 101 - 107 is connected by an internal bus 110 .

The CPU 101 controls the biometric information arithmetic device 1 as a whole. ROM 102 stores the operation code of CPU 101 . A RAM 103 is a work area used when the CPU 101 operates. The storage unit 104 stores various information such as databases and evaluation data. As the storage unit 104, for example, a data storage device such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive) is used. In addition, for example, the biological information arithmetic device 1 may have a GPU (Graphics Processing Unit) not shown.

The I/F 105 is an interface for transmitting/receiving various kinds of information to/from the server 4, the sensor 5, etc. via the communication network 3 as necessary. The I/F 106 is an interface for transmitting/receiving information to/from the input unit 108 . A keyboard, for example, is used as the input unit 108 , and the user or the like of the biological information computing device 1 inputs various kinds of information, control commands for the biological information computing device 1 , etc. via the input unit 108 . The I/F 107 is an interface for transmitting and receiving various information to and from the display unit 109 . The display unit 109 displays various information stored in the storage unit 104, evaluation results, and the like. A display is used as the display unit 109 , and is provided integrally with the input unit 108 in the case of a touch panel type, for example.

FIG. 4B is a schematic diagram showing an example of the functions of the biological information computing device 1. As shown in FIG. The biological information computing device 1 includes an acquisition unit 11, a generation unit 12, an evaluation unit 13, an output unit 14, and a storage unit 15, and may include a learning unit 16, for example. Each function shown in FIG. 4B is realized by the CPU 101 using the RAM 103 as a work area and executing a program stored in the storage unit 104 or the like.

<Acquisition unit 11>
Acquisition unit 11 acquires evaluation data based on a user's pulse wave. For example, after acquiring sensor data from the sensor 5 or the like, the acquiring unit 11 acquires evaluation data by processing the sensor data.

For example, as shown in FIG. 3B, the acquisition unit 11 performs filtering processing on the acquired sensor data. A bandpass filter of 0.5 to 5.0 Hz, for example, is used in the filtering process. Thereby, the acquiring unit 11 extracts data (pulse wave data) corresponding to the user's pulse wave. The pulse wave data indicates data corresponding to velocity pulse wave, for example. The pulse wave data may indicate data corresponding to an acceleration pulse wave or a volume pulse wave, for example, and can be arbitrarily set according to the type and application of the sensor. Also, the filter range of the bandpass filter can be arbitrarily set according to the application.

Acquisition unit 11 performs differentiation processing on, for example, pulse wave data. For example, when differential processing is performed on pulse wave data corresponding to a velocity pulse wave, the acquiring unit 11 acquires data (differential data) corresponding to an acceleration pulse wave. Note that in the differentiation process, two differentiations may be performed in addition to the one-time differentiation.

The acquisition unit 11 performs division processing on, for example, differential data. In the division process, for example, the differential data corresponding to multiple cycles of the acceleration pulse wave is divided into data (divided data) corresponding to the acceleration pulse wave for each cycle. Therefore, the obtaining unit 11 can obtain a plurality of pieces of divided data by, for example, performing differentiation processing on one piece of differential data. Note that in the division process, the differential data can be divided for every arbitrary period (for example, an integer multiple of the period) depending on the application.

For example, in the division process, the data amount of each divided data may be different. In this case, the acquisition unit 11 may specify the divided data with the smallest data amount, and reduce the data amount (trimming) of the other divided data. This makes it possible to unify the amount of data in each piece of divided data, making it easier to compare data in each piece of divided data.

In addition to the above, for example, normalization processing may be performed on values corresponding to the time axis of divided data. In the normalization process, for example, normalization is performed by setting 0 as the minimum value and 1 as the maximum value of the values corresponding to the time axis. This facilitates comparison of data in each divided data.

For example, the acquisition unit 11 may reduce the data amount or calculate the average of a plurality of pieces of divided data that have been normalized, and use the average as divided data.

Acquisition unit 11 performs normalization processing on the divided data. In the normalization process, normalized data (normalized data) is generated for values corresponding to amplitudes. In the normalization process, for example, normalization is performed by setting 0 to the lowest value of amplitude and 1 to the highest value of amplitude. The acquisition unit 11 acquires, for example, normalized data as evaluation data. In this case, data corresponding to the acceleration pulse wave of the user is obtained as the evaluation data.

The acquiring unit 11 may sequentially perform the above-described processes, and may not perform differentiation as shown in FIG. 3C, for example. In this case, data corresponding to the user's velocity pulse wave is obtained as the evaluation data.

Further, the acquisition unit 11 may perform only a part of each process described above, for example. In this case, the acquisition unit 11 may acquire, as evaluation data, any of the pulse wave data, the differential data, the divided data, the trimmed divided data, and the divided data obtained by normalizing the values corresponding to the time axis. , can be set arbitrarily according to the application.

Note that the acquisition unit 11 may acquire evaluation target information such as the user's characteristics, exercise content, sporting event, etc., input by the user via the input unit 108 or the like, and include the information in the evaluation data. The evaluation target information may be used, for example, when generating a blood sugar level evaluation result or an evaluation result.

<Generating unit 12>
The generation unit 12 refers to the database and generates a blood glucose level evaluation result for the evaluation data. The generation unit 12 refers to, for example, classification information stored in a database, calculates a blood sugar level for the evaluation data, and generates a blood sugar level evaluation result. The generation unit 12 generates a plurality of blood glucose level evaluation results for different evaluation data.

<Evaluation unit 13>
The evaluation unit 13 generates an evaluation result based on the evaluation data and the blood sugar level evaluation result generated from the evaluation data. The evaluation unit 13 uses a display format stored in advance in the storage unit 104 or the like, for example, to generate an evaluation result of the health condition converted into a format that the user can understand.

The evaluation unit 13 may refer to, for example, an evaluation database to generate an evaluation result from the evaluation data and the blood glucose level evaluation result.

The evaluation database may store evaluation classification information for generating evaluation results for evaluation data and blood glucose level evaluation results, for example, in the same manner as the database described above. In addition to storing one or more pieces of evaluation classification information, the evaluation database may store, for example, a plurality of pieces of evaluation learning data used to generate the evaluation classification information.

The evaluation classification information is, for example, a function indicating the correlation between past evaluation data and blood glucose level evaluation results (evaluation input data) obtained in advance and evaluation reference data linked to the evaluation input data. The evaluation reference data indicates user evaluation results. The classification information for evaluation is generated using a plurality of learning data for evaluation, with the input data for evaluation and the reference data for evaluation as a pair of learning data for evaluation.

The classification information for evaluation indicates a calibration model generated based on the results of the above-described regression analysis using, for example, the input data for evaluation as explanatory variables and the reference data for evaluation as objective variables. For example, the evaluation classification information can periodically update the calibration model (evaluation calibration model), and also can be classified according to the user's gender, age, exercise content, physique, BMI (Body Mass Index), and other attribute information. may be generated. Note that the classification information for evaluation may include a trained model (learned model for evaluation) generated by machine learning using a plurality of learning data for evaluation, for example, in the same manner as the classification information described above.

<Output unit 14>
The output unit 14 outputs evaluation results. In addition to outputting the evaluation result to the display unit 109, the output unit 14 may output the evaluation result to the sensor 5, for example.

<Storage unit 15>
The storage unit 15 retrieves various data such as databases stored in the storage unit 104 as necessary. The storage unit 15 stores various data acquired or generated by each of the components 11 to 14 and 16 in the storage unit 104 as necessary.

<Learning part 16>
The learning unit 16 generates classification information using, for example, a plurality of learning data. The learning unit 16 may acquire new data for learning, for example, and update the existing classification information.

The learning unit 16 generates evaluation classification information using, for example, a plurality of evaluation learning data. The learning unit 16 may, for example, acquire new learning data for evaluation and update the existing classification information for evaluation.

When the classification information and the evaluation classification information are used in the biological information computing system 100, the classification information may be updated according to the type of evaluation data, and the evaluation classification information may not be updated. In this case, since there is no need to newly prepare learning data for evaluation, it is possible to realize a significant reduction in cost and update time.

<Communication network 3>
The communication network 3 is a known Internet network or the like that connects the biological information computing device 1, the server 4, and the sensor 5 via communication lines. The communication network 3 may be configured by a LAN (Local Area Network) or the like when the biometric information computing system 100 is operated within a fixed narrow area. Moreover, the communication network 3 may be configured by a so-called optical fiber communication network. Moreover, the communication network 3 is not limited to a wired communication network, and may be realized by a wireless communication network, and can be arbitrarily set according to the application.

The server 4 accumulates information sent via the communication network 3 . The server 4 transmits the accumulated information to the biometric information computing device 1 via the communication network 3 based on the request from the biometric information computing device 1 .

<Sensor 5>
The sensor 5 generates sensor data. The sensor 5 includes a detection section 6, for example, as shown in FIG. 5(a). The sensor 5 is attached to a position where the user's pulse wave can be detected via the detection unit 6, and is fixed to a wristband 55, for example.

A known detection device capable of detecting a user's pulse wave is used as the detection unit 6 . As the detection unit 6, for example, a strain sensor such as a fiber Bragg grating (FBG) sensor, a gyro sensor, one or more electrodes for pulse wave signal measurement, a photoplethysmographic pulse wave (PPG) sensor, a pressure sensor, and an optical detection module. is used. A plurality of detection units 6 may be arranged, for example.

Note that the sensor 5 may be embedded in clothing. In addition to humans, users who wear the sensor 5 may target pets such as dogs and cats, and may target livestock such as cattle and pigs, and farming of fish and the like.

For example, as shown in FIG. 5B, the sensor 5 includes an acquisition unit 50, a communication I/F 51, a memory 52, and an instruction unit 53, each of which is connected via an internal bus .

The acquisition unit 50 measures the user's pulse wave via the detection unit 6 and generates sensor data. The acquisition unit 50 transmits the generated sensor data, for example, to the communication I/F 51 or the memory 52 .

The communication I/F 51 transmits various data such as sensor data to the biological information computing device 1 and the server 4 via the communication network 3 . In addition, the communication I/F 51 includes a line control circuit for connecting to the communication network 3, a signal conversion circuit for performing data communication with the biological information arithmetic device 1 and the server 4, and the like. The communication I/F 51 converts various commands from the internal bus 54 and sends them to the communication network 3 side. to the internal bus 54.

The memory 52 stores various data such as sensor data transmitted from the acquisition unit 50 . The memory 52 transmits various data such as stored sensor data to the communication I/F 51 by receiving a command from another terminal device connected via the communication network 3, for example.

The instruction unit 53 includes operation buttons, a keyboard, and the like for acquiring sensor data, and includes a processor such as a CPU, for example. The instruction unit 53 notifies the acquisition unit 50 of the received instruction to acquire the sensor data. The acquiring unit 50 that has received this notification acquires the sensor data. Note that the instruction unit 53 may perform processing for acquiring evaluation data from sensor data, as shown in FIGS. 3B and 3C, for example.

Here, a case of using an FBG sensor will be described as an example of acquiring sensor data.

The FBG sensor has a diffraction grating structure formed at predetermined intervals in one optical fiber. The FBG sensor has, for example, a sensor portion length of 10 mm, a wavelength resolution of ±0.1 pm, a wavelength range of 1550±0.5 nm, a fiber diameter of 145 μm, and a core diameter of 10.5 μm. Measurement can be performed while the FBG sensor is in contact with the skin of the user as the detection unit 6 described above.

For example, an ASE (Amplified Spontaneous Emission) light source with a wavelength range of 1525 to 1570 nm is used as a light source for optical fibers. Emitted light from the light source enters the FBG sensor via the circulator. Reflected light from the FBG sensor is guided to a Mach-Zehnder interferometer via a circulator, and output light from the Mach-Zehnder interferometer is detected by a photodetector. A Mach-Zehnder interferometer is used to separate two optical paths with an optical path difference by a beam splitter and combine them again by a beam splitter to produce interference light. For example, the length of one optical fiber may be increased to provide an optical path difference. Coherent light produces interference fringes in accordance with the optical path difference, so by measuring the pattern of the interference fringes, it is possible to detect a change in strain, ie, a pulse wave, occurring in the FBG sensor. Acquisition unit 50 generates sensor data based on the detected pulse wave. Thereby, sensor data is acquired.

In addition, the optical fiber sensor system that detects the distortion amount of the FBG sensor and detects the waveform of the pulse wave includes a wide band ASE light source, a circulator, a Mach-Zehnder interferometer, and a beam splitter in addition to the light source that is incident on the FBG sensor. optical system, a light receiving sensor included in the photodetector, and an analysis method for analyzing the amount of wavelength shift. The optical fiber sensor system can be used by selecting a light source and band light according to the characteristics of the FBG sensor to be used, and various methods of analysis such as a detection method can be adopted.

(First embodiment: operation of biological information computing system 100)
Next, an example of the operation of the biological information computing system 100 according to this embodiment will be described. FIG. 6 is a flow chart showing an example of the operation of the biological information computing system 100 according to this embodiment.

The biometric information computing system 100 executes, for example, via a biometric information computing program installed in the biometric information computing device 1 . That is, the user operates the biometric information computing device 1 or the sensor 5, and obtains an evaluation result indicating the health condition of the user from the sensor data through the biometric information computing program installed in the biometric information computing device 1. can be done.

The operation of the biological information computing system 100 includes an acquisition step S110, a generation step S120, an evaluation step S130, and may include an output step S140, for example.

<Acquisition step S110>
Acquisition step S110 acquires evaluation data based on the user's pulse wave. For example, the acquisition unit 50 of the sensor 5 measures the user's pulse wave via the detection unit 6 and generates sensor data. Acquisition unit 50 transmits the sensor data to biological information calculation device 1 via communication I/F 51 and communication network 3 . The acquisition unit 11 of the biological information computing device 1 receives sensor data from the sensor 5 .

The acquisition unit 11 performs, for example, the process shown in FIG. 3B on sensor data to acquire evaluation data. The acquisition unit 11 stores the acquired evaluation data in the storage unit 104 via the storage unit 15, for example. Conditions such as the frequency with which the acquisition unit 11 acquires sensor data from the sensor 5 can be arbitrarily set according to the application. For example, the acquisition unit 11 acquires evaluation data at a preset cycle. In this case, it is possible to simplify the arithmetic processing when generating the evaluation result, so that it is possible to improve the processing speed.

<Generation Step S120>
Next, a generating step S120 refers to the database and generates a blood glucose level evaluation result including the blood glucose level for the evaluation data. For example, the generation unit 12 refers to the classification information and calculates the blood sugar level for the evaluation data. The generation unit 12 generates a blood sugar level evaluation result including the calculated value. The generation unit 12 stores the generated blood glucose level evaluation result in the storage unit 104 via the storage unit 15, for example. As the blood sugar level, in addition to indicating a specific value, for example, an error range (for example, "○○±5 mg/dL", etc.) may be calculated.

<Evaluation step S130>
Next, an evaluation step S130 generates an evaluation result based on the evaluation data and the blood glucose level evaluation result generated from the evaluation data. For example, the evaluation unit 13 acquires the evaluation data and the blood glucose level evaluation result generated by the generation unit 12 . The evaluation unit 13 generates an evaluation result from the evaluation data and the blood glucose level evaluation result using a preset function or the like, and may generate the evaluation result by referring to the evaluation database described above, for example.

The evaluation unit 13 generates an evaluation result by, for example, combining the blood glucose level evaluation result with the evaluation data. In this case, in addition to using a preset function or the like, an evaluation database may be referenced to generate an evaluation result. This makes it possible to easily consider the characteristics of the user's pulse wave and improve the accuracy of the evaluation result.

<Output step S140>
Next, for example, an output step S140 may output the evaluation result. For example, the output unit 14 outputs evaluation results to the display unit 109 .

Thus, the operation of the biometric information computing system 100 ends. The frequency and order of performing each step can be arbitrarily set according to the application.

According to this embodiment, the generation unit 12 generates a blood sugar level evaluation result including a blood sugar level for the evaluation data. The evaluation unit 13 generates an evaluation result regarding the health condition of the user based on the evaluation data and the blood sugar level evaluation result generated from the evaluation data. Therefore, the evaluation result can be generated from the evaluation data obtained based on the pulse wave characteristics and the blood glucose level evaluation result. This makes it possible to generate user evaluation results with high accuracy.

Further, according to the present embodiment, the generation unit 12 refers to the database and generates the blood glucose level evaluation result for the evaluation data. The database also stores classification information calculated using a plurality of pieces of learning data. Therefore, when generating an evaluation result, it is possible to include a quantitative blood sugar level based on data that has been proven in the past. This makes it possible to improve the accuracy when generating evaluation results.

Further, according to the present embodiment, the classification information is a calibration model obtained using PLS regression analysis with input data as explanatory variables and reference data as objective variables. Therefore, compared to the case of calculating classification information using machine learning or the like, it is possible to greatly reduce the number of data for learning and to easily update the calibration model. This makes it possible to facilitate construction and updating of the biometric information computing system 100 .

(Second embodiment: biological information computing system 100)
Next, an example of the biometric information computing system 100 in the second embodiment will be described. The difference between the above-described embodiment and the second embodiment is that a function for generating an evaluation result from the blood glucose level evaluation result is selected based on the characteristics of the evaluation data, and the evaluation result is generated from the blood glucose level evaluation result. is. In addition, description is abbreviate|omitted about the content similar to embodiment mentioned above.

The evaluation unit 13 may determine a calculation method for the blood glucose level evaluation result, for example, according to the characteristics of the evaluation data. In this case, a function or the like for generating an evaluation result from different blood glucose evaluation results for each feature of the evaluation data is stored in an evaluation database stored in the storage unit 104 or the like, for example. The evaluation unit 13 may select a function stored in the storage unit 104 based on features such as peak intensity and shape in the evaluation data, and generate an evaluation result from the blood glucose level evaluation result. For example, based on the characteristics of the evaluation data, the evaluation unit 13 may select, as the function described above, evaluation classification information suitable for the evaluation data from a plurality of evaluation classification information.

For example, when data corresponding to a user's acceleration pulse wave is used as evaluation data, the "feature of evaluation data" is determined by whether it belongs to one of the seven classification patterns (A to G) shown in FIG. may be In addition, for example, when data corresponding to the user's velocity pulse wave is used as the evaluation data, the "feature of the peak of the evaluation data" is classified into one of two types (group 1 and group 2) as shown in FIG. It may be determined by belonging to a pattern.

The evaluation database may store evaluation classification information for generating evaluation results for blood glucose level evaluation results, for example, in the same manner as the database described above. In addition to storing one or more pieces of evaluation classification information, the evaluation database may store, for example, a plurality of pieces of evaluation learning data used to generate the evaluation classification information.

The evaluation classification information is, for example, a function indicating the correlation between the blood glucose level evaluation result (evaluation input data) linked to past evaluation data acquired in advance and the evaluation reference data linked to the evaluation input data. . The evaluation reference data indicates user evaluation results. The classification information for evaluation is generated using a plurality of learning data for evaluation, with the input data for evaluation and the reference data for evaluation as a pair of learning data for evaluation.

In the biological information computing system 100 of the present embodiment, for example, as shown in FIG. 9, the evaluation step S130 includes an evaluation selection step S131 and an attribute-based evaluation step S132.

The evaluation selection step S131 refers to the evaluation data and selects specific evaluation classification information (for example, first evaluation classification information) from among the plurality of attribute-based evaluation classification information. The evaluation selection step S131 can be executed by an evaluation selection unit included in the evaluation unit 13, for example.

The attribute-based evaluation step S132 refers to the selected first evaluation classification information and generates an evaluation result for the blood sugar level evaluation result. The attribute-based evaluation step S132 can be executed by an attribute-based evaluation unit included in the evaluation unit 13, for example.

A plurality of attribute-based evaluation classification information includes different functions for generating evaluation results from blood glucose level evaluation results. Also, the plurality of attribute-based evaluation classification information may be calculated using learning data. For example, as input data for learning data, blood sugar level evaluation results linked to the evaluation data for each of the classification patterns described above are used. In this case, for example, input data is prepared for each blood glucose level evaluation result linked to evaluation data of seven types (A to G) of classification patterns as shown in FIG. 7, and seven types of attribute-based evaluation classification information are generated. .

When such a plurality of attribute-based evaluation classification information are stored in the database, for example, the acquisition unit 11 acquires evaluation data corresponding to the user's acceleration pulse wave. Then, the evaluation unit 13 refers to the characteristics of the peaks of the evaluation data and selects the first classification information for evaluation. After that, the evaluation unit 13 refers to the first evaluation classification information and generates an evaluation result for the blood sugar level evaluation result. Therefore, it is possible to select the most suitable evaluation classification information for the user from among the attribute-based evaluation classification information.

For example, when data corresponding to the velocity pulse wave of the subject is used as input data for learning data, for example, two types (group 1, group 2) of classification pattern evaluation data as shown in FIG. Input data may be prepared for each linked blood glucose level evaluation result, and two types of attribute-based evaluation classification information may be generated. This makes it possible to further improve the evaluation accuracy.

Here, an example of data used for the classification patterns described above will be described.

For example, as shown in FIG. 7, the acceleration pulse wave has inflection points a to e. For example, the maximum peak in the acceleration pulse wave is a point, each inflection point is b point, c point, d point, e point in order from a point, a point is 1, and the minimum value is b point or d When normalization is performed with the point set to 0, the acceleration pulse wave can be classified into 7 patterns using a method of classifying according to the value of each inflection point and the magnitude relationship between the differences. First, when the value of the inflection point is b<d, it is classified into pattern A or B. FIG. If b<d and c≧0.5, it is classified as A; otherwise, it is classified as B. Next, when the value of the inflection point is b≈d, it is classified into pattern C or D. If b ≈ d and c ≈ 0, then it is classified as pattern D; otherwise, it is classified as pattern C. Finally, if b>d, it can be classified into one of patterns E, F, and G. If b>d and b<c, it is classified into pattern E. If b≈c, it is classified into pattern F. If b>c, it is classified into pattern G.

For example, the evaluation unit 13 determines, for example, which pattern in FIG. 7 the evaluation data applies to, and identifies the classification pattern. For example, if the inflection point b of the input evaluation data is smaller than the inflection point d and the inflection point c≧0.5, the pattern A is set as the classification pattern. This makes it possible to refer to the evaluation classification information suitable for the characteristics of the evaluation data and accurately calculate the evaluation result.

According to this embodiment, in addition to the effects of the above-described embodiments, the evaluation unit 13 selects a function for generating evaluation results from blood glucose level evaluation results based on the characteristics of evaluation data. Therefore, an appropriate function can be selected according to the characteristics of the evaluation data that differ from user to user. This makes it possible to generate user evaluation results with high accuracy.

(Third embodiment: biological information computing system 100)
Next, an example of the biometric information computing system 100 in the third embodiment will be described. The difference between the above-described embodiment and the third embodiment is that a blood sugar trend step S150 is provided. In addition, description is abbreviate|omitted about the content similar to embodiment mentioned above.

In the biological information computing system 100 according to the present embodiment, for example, the evaluation unit 13 described above generates a blood sugar trend based on a plurality of blood sugar level evaluation results generated from different evaluation data. Also, the evaluation unit 13 generates an evaluation result from the evaluation data and the blood sugar trend. The blood sugar trend is, for example, a graph showing changes in blood sugar level with respect to time t, as shown in FIG. Also, the blood sugar trend may be obtained by differentiating or integrating a graph showing changes in blood sugar level with respect to time t, for example, with respect to time t. The blood glucose trend may also include blood glucose spikes that indicate the maximum blood glucose level. The blood sugar trend may also include the HbA1c value.

The evaluation unit 13 acquires, for example, a plurality of blood sugar level evaluation results generated from different evaluation data, and acquires a blood sugar trend related to the tendency of the user's blood sugar level from time-series changes in the blood sugar level evaluation results.

The evaluation unit 13 may refer to, for example, an evaluation database to generate an evaluation result from the evaluation data and the blood sugar trend.

The evaluation database may store evaluation classification information for generating evaluation data and evaluation results for blood sugar trends, for example, in the same manner as the database described above. In addition to storing one or more pieces of evaluation classification information, the evaluation database may store, for example, a plurality of pieces of evaluation learning data used to generate the evaluation classification information.

The evaluation classification information is, for example, a function indicating the correlation between past evaluation data and blood glucose trends (evaluation input data) obtained in advance and evaluation reference data linked to the evaluation input data. The evaluation reference data indicates user evaluation results. The classification information for evaluation is generated using a plurality of learning data for evaluation, with the input data for evaluation and the reference data for evaluation as a pair of learning data for evaluation.

(Third Embodiment: Operation of biological information computing system 100)
Next, an example of the operation of the biological information computing system 100 according to this embodiment will be described. FIG. 11 is a flow chart showing an example of the operation of the biological information computing system 100 in this embodiment. As for the operation of the biological information computing system 100 in this embodiment, as shown in FIG. 11, the evaluation step S130 includes the blood sugar trend step S150.

<Blood Sugar Trend Step S150>
The blood sugar trend step S150 acquires a plurality of blood sugar level evaluation results generated from, for example, different evaluation data, and acquires a blood sugar trend related to the tendency of the user's blood sugar level from time-series changes in the blood sugar level evaluation results. The blood sugar trend step S150 can be executed by the evaluation unit 13 included in the biological information computing device 1, for example.

<Evaluation step S130>
Next, an evaluation step S130 generates an evaluation result based on the evaluation data and the blood glucose trend. For example, the evaluation unit 13 acquires the evaluation data and the blood sugar trend generated by the generation unit 12 . The evaluation unit 13 generates an evaluation result from the evaluation data and the blood sugar trend using a preset function or the like, and may generate the evaluation result by referring to the evaluation database described above, for example. Therefore, it is possible to easily grasp characteristics of blood sugar levels, such as blood sugar level spikes, which are different for each user. This makes it possible to generate user evaluation results with high accuracy.

(Fourth embodiment: biological information computing system 100)
Next, an example of the biometric information computing system 100 in the fourth embodiment will be described. The difference between the above-described embodiment and the fourth embodiment is that additional information is used. In addition, description is abbreviate|omitted about the content similar to embodiment mentioned above.

In the evaluation step S130, for example, as shown in FIG. 12, additional information is acquired and an evaluation result is generated based on the evaluation data, the blood glucose level evaluation result, and the additional information. The evaluation step S130 may also include a blood glucose trend step S150. In this case, the evaluation step S130 acquires the additional information and generates an evaluation result based on the evaluation data, the blood sugar trend and the additional information.

The additional information indicates characteristics of the user and includes, for example, at least one of the user's stress level, pulse rate, breathing rate, blood pressure, and blood sugar level.

As additional information, for example, data measured using a known measurement method is used, and the data format is arbitrary. Further, the timing of acquiring the additional information may be any timing other than the timing of measuring the pulse wave, depending on the application.

The additional information includes, for example, attribute information such as the user's gender and age, and may also include information related to exercise at the time of evaluation, such as the details of the exercise the user is doing and the type of competition. The additional information is input by the user through the input unit 108 or the like, and is acquired by the evaluation unit 13 or the like.

The evaluation unit 13 may refer to an evaluation database, for example, and generate an evaluation result suitable for the evaluation data, the blood glucose level evaluation result, and the additional information.

The evaluation database may store evaluation classification information for generating evaluation data, blood glucose level evaluation results, and evaluation results for additional information, for example, in the same manner as the database described above.

The evaluation classification information is, for example, a function that indicates the correlation between past evaluation data, blood glucose level evaluation results, and additional information (evaluation input data) obtained in advance, and evaluation reference data linked to the evaluation input data. be. The evaluation reference data indicates user evaluation results. The classification information for evaluation is generated using a plurality of learning data for evaluation, with the input data for evaluation and the reference data for evaluation as a pair of learning data for evaluation.

The classification information for evaluation indicates a calibration model generated based on the results of the above-described regression analysis using, for example, the input data for evaluation as explanatory variables and the reference data for evaluation as objective variables. The classification information for evaluation can periodically update the calibration model (calibration model for evaluation), for example, or may be generated for each additional information, for example. Note that the classification information for evaluation may include a trained model (learned model for evaluation) generated by machine learning using a plurality of learning data for evaluation, for example, in the same manner as the classification information described above.

According to the present embodiment, in addition to the effects of the above-described embodiments, the evaluation unit 13 acquires additional information indicating user characteristics, and generates an evaluation result based on the evaluation data, the blood glucose level evaluation result, and the additional information. Including. Therefore, by using a plurality of types of information, it is possible to realize comprehensive evaluation based on multifaceted viewpoints. This makes it possible to generate user evaluation results with higher accuracy.

(Fifth embodiment: biological information computing system 100)
Next, an example of the biometric information computing system 100 according to the fifth embodiment will be described. The difference between the above-described embodiment and the fifth embodiment is that additional data for generating additional information is acquired. In addition, description is abbreviate|omitted about the content similar to embodiment mentioned above.

In the biometric information computing system 100 of this embodiment, the obtaining step S110 includes obtaining additional data. In addition, in the biometric information computing system 100 of this embodiment, the generating step S120 includes generating additional information based on additional data. The evaluation step S130 may also include a blood glucose trend step S150. In this case, the evaluation step S130 generates an evaluation result based on the evaluation data, blood sugar trend and additional information.

For example, as shown in FIG. 13, the biological information computing device 1 in this embodiment performs two types of processing on one sensor data generated based on the pulse wave. Thereby, for example, the acquiring unit 11 acquires the evaluation data and the additional data indicating characteristics different from the evaluation data based on the same pulse wave. Note that a plurality of pieces of additional data may be acquired.

For example, the generator 12 generates additional information based on additional data. In this embodiment, for example, at least one of the pulse rate and respiration rate can be calculated as the additional information. In this case, for example, as shown in FIGS. 14A and 14B, the contents of the preprocessing performed in the acquisition step S110 and the contents of the database referred to in the generation step S120 may differ. An example of calculating the pulse rate and respiration rate as additional information will be described below.

<Calculation of pulse rate>
For example, when calculating the pulse rate, as shown in FIG. 14A, the acquiring unit 11 performs the filtering process described above on the sensor data to extract pulse wave data.

Then, the acquisition unit 11 performs peak position calculation processing on the pulse wave data. In the peak position calculation process, a plurality of peaks (maximum amplitude values) included in the pulse wave data are detected, and the sampling order (corresponding to the time from the start of measurement) is specified. Thereby, the acquisition unit 11 acquires the peak position data included in the pulse wave data.

After that, the acquisition unit 11 performs peak interval average calculation processing on the peak position data. The peak interval average calculation process calculates intervals between peaks included in the peak position data (differences in the order in which adjacent peaks are sampled), and calculates, for example, an average value of the peak intervals. After that, the acquisition unit 11 divides the peak interval or the average value of the peak intervals by the sampling rate of the sensor data, and acquires data indicating the peak interval corresponding to the number of seconds as additional data.

After that, the generator 12 refers to the database and calculates the pulse rate for the additional data. At this time, the generator 12 refers to the pulse rate classification information stored in the database. The pulse rate classification information indicates, for example, a function that divides 60 [seconds] by the peak interval. Therefore, the generator 12 can calculate the pulse rate (=71 [bpm]) for the additional data (peak interval=0.85 [seconds]), for example. Thereby, the generator 12 can generate additional information including the pulse rate.

<Calculation of respiratory rate>
For example, when calculating the respiratory rate, as shown in FIG. 14B, the acquiring unit 11 performs the filtering process described above on the sensor data and extracts the pulse wave data.

After that, the acquiring unit 11 performs Fourier transform processing on the pulse wave data. In Fourier transform processing, for example, pulse wave data representing sampling time versus amplitude is transformed into frequency data representing frequency versus intensity. Thereby, the acquiring unit 11 acquires the frequency data for the pulse wave data.

After that, the acquisition unit 11 performs maximum frequency detection processing on the frequency data. In the maximum frequency detection process, the maximum intensity frequency between 0.15 and 0.35 Hz is specified in the frequency data. Thereby, the acquiring unit 11 acquires the specified frequency value as the additional data.

After that, the generation unit 12 refers to the database and calculates the respiration rate for the additional data. At this time, the generation unit 12 refers to the respiratory rate classification information stored in the database. The respiratory rate classification information indicates, for example, a function of multiplying the specified frequency by 60 [seconds]. Therefore, the generation unit 12 can calculate, for example, the respiration rate (=13.5 [bpm]) for the additional data (specified frequency=0.225 Hz). Thereby, the generator 12 can generate additional information including, for example, the respiration rate.

In this manner, the biological information computing device 1 can set preprocessing for sensor data according to the content of biological information to be included in the additional information, and can arbitrarily set the content of the database to be referred to.

In addition to the above-described pulse rate and respiration rate, information that can be calculated based on pulse waves, such as blood sugar level, blood pressure, blood vessel age, degree of diabetes, and the like, can be used as biological information. The biometric information may indicate, for example, lactate levels.

Further, classification information (preliminary classification information) used for calculation may be stored in the database in advance according to the content of the biometric information to be calculated. In this case, a plurality of pairs of preliminary input data indicating the learning pulse wave and preliminary reference data including biological information linked to the preliminary input data are prepared as preliminary learning data. Then, for example, the learning unit 16 generates preliminary classification information using a plurality of pieces of preliminary learning data, and stores it in the database.

The biometric information contained in the preliminary reference data can be obtained by measuring the subject using a measuring device for measuring the necessary biometric information, for example, in the same manner as the reference data described above. . Also, as a method for learning the preliminary classification information, the same method as for the classification information described above can be used.

The biological information computing system 100 in this embodiment may output the generated additional information. In this case, it is possible to use additional information generated from a viewpoint different from the blood sugar level, and it is possible to generate a multifaceted evaluation result that considers the characteristics of the user. This makes it possible to generate user evaluation results with higher accuracy.

According to the present embodiment, in addition to the effects of the above-described embodiments, the acquisition unit 11 acquires additional data indicating features different from the evaluation data based on the pulse wave. The generation unit 12 also generates additional information based on the additional data. That is, the evaluation result is generated from the evaluation data generated based on one pulse wave, the blood sugar level evaluation result, and the additional information. Therefore, by using a plurality of types of information based on the same parameter, it is possible to realize a comprehensive evaluation based on multiple viewpoints. This makes it possible to generate user evaluation results with higher accuracy.

(Sixth embodiment: biological information computing system 100)
Next, an example of the biometric information computing system 100 according to the sixth embodiment will be described. The difference between the above-described embodiment and the sixth embodiment is that the above-described additional information is acquired in the generation step S120, and the blood glucose level evaluation result is generated based on the evaluation data and the additional information. In addition, description is abbreviate|omitted about the content similar to embodiment mentioned above.

In the biological information computing system 100 of this embodiment, for example, as shown in FIG. 15, the generation step S120 includes obtaining additional information and generating a blood glucose level evaluation result based on the evaluation data and the additional information. The evaluation step S130 may also include a blood glucose trend step S150. In this case, the evaluation step S130 generates an evaluation result based on the evaluation data and the blood glucose trend.

The additional information has the same content as described above, and is input by the user via the input unit 108 or the like and acquired by the generation unit 12 or the like. For example, the generator 12 may acquire the additional information generated by the generator 12 based on the additional data acquired by the acquirer 11 .

The generation unit 12 may determine the calculation method for the blood glucose level evaluation result, for example, according to the content of the additional information. In this case, functions and the like that differ for each type of additional information are stored in the storage unit 104 and the evaluation database. Note that the generating unit 12 may generate the blood glucose level evaluation result based on information obtained by combining evaluation data and additional information, for example.

According to this embodiment, in addition to the effects of the above-described embodiments, the generator 12 acquires additional information and generates a blood glucose level evaluation result based on the evaluation data and the additional information. For this reason, it is possible to use a parameter correlated with the user's blood sugar level evaluation result and realize a comprehensive evaluation based on multiple viewpoints. This makes it possible to generate user evaluation results with higher accuracy.

(Seventh embodiment: biological information computing system 100)
Next, an example of the biometric information computing system 100 according to the seventh embodiment will be described. The difference between the above-described embodiment and the seventh embodiment is that classification information suitable for evaluation data is selected from a plurality of classification information based on additional information. In addition, description is abbreviate|omitted about the content similar to embodiment mentioned above.

In the biological information computing device 1 according to the present embodiment, for example, as shown in FIG. 16, the generation step S120 includes a selection step S121 and an attribute-based generation step S122. The evaluation step S130 may also include a blood glucose trend step S150. In this case, the evaluation step S130 generates an evaluation result based on the evaluation data and the blood glucose trend.

A selection step S121 selects specific classification information (for example, first classification information) from a plurality of attribute-based classification information with reference to the additional information. The selection step S121 can be executed by a selection unit included in the generation unit 12, for example.

In the attribute-by-attribute generation step S122, the biological information computing device 1 refers to the selected first classification information, calculates the blood sugar level for the evaluation data, and generates the blood sugar level evaluation result. The attribute-based generation step S<b>122 can be executed by an attribute-based generation unit included in the generation unit 12 , for example.

A plurality of pieces of attribute-based classification information are calculated using different learning data. For example, as input data for learning data, input data is prepared for each feature of additional information to generate a plurality of attribute-based classification information. For example, when the additional information is the user's age, input data may be prepared for each age group such as 10 to 30 years old, 30 to 50 years old, 50 to 70 years old, and the like.

When such a plurality of attribute-based classification information are stored in the database, for example, the acquisition unit 11 acquires evaluation data and additional information. Then, the generation unit 12 refers to the additional information and selects the first classification information. After that, the generating unit 12 refers to the first classification information and generates a blood glucose level evaluation result for the evaluation data. Therefore, it is possible to select the most suitable classification information for the user from among the respective attribute-based classification information.

In this case, as the obtaining step S110, for example, the obtaining unit 11 obtains the user's evaluation data and the age of the user as additional information.

Next, as a selection step S121, for example, the generation unit 12 refers to the additional information and selects the first classification information associated with the additional information. Then, as an attribute-based generation step S122, the generation unit 12 refers to the first classification information and generates a blood glucose level evaluation result for the evaluation data. This makes it possible to further improve the evaluation accuracy.

According to the present embodiment, in addition to the effects of the above-described embodiments, the generation unit 12 refers to the additional information and selects the first classification information, refers to the first classification information, and an attribute-by-attribute generator for generating a value evaluation result. Therefore, it is possible to select the most suitable attribute-based classification information for the additional information and then generate the blood glucose level evaluation result for the evaluation data. This makes it possible to further improve the evaluation accuracy.

(Eighth embodiment: biological information computing system 100)
Next, an example of the biometric information computing system 100 in the eighth embodiment will be described. The difference between the above-described embodiment and the eighth embodiment is that preliminary evaluation data is referred to and classification information suitable for evaluation data is selected from a plurality of classification information. In addition, description is abbreviate|omitted about the content similar to embodiment mentioned above.

In the biometric information computing device 1 of the present embodiment, for example, as shown in FIG. 17, the generation step S120 includes a selection step S121 and an attribute-based generation step S122.

A selection step S121 refers to preliminary evaluation data and selects specific classification information (for example, first classification information) from a plurality of attribute-based classification information. The selection step S121 can be executed by a selection unit included in the generation unit 12, for example.

The preliminary evaluation data exhibits characteristics different from those of the evaluation data, such as characteristics similar to those of the additional data described above. The preliminary evaluation data may be used, for example, to generate additional information, similar to the additional data described above.

In the attribute-by-attribute generation step S122, the biological information computing device 1 refers to the selected first classification information, calculates the blood sugar level for the evaluation data, and generates the blood sugar level evaluation result. The attribute-based generation step S<b>122 can be executed by an attribute-based generation unit included in the generation unit 12 , for example.

A plurality of pieces of attribute-based classification information are calculated using different learning data. For example, when data corresponding to the acceleration pulse wave of the subject is used as input data for learning data, input data is prepared for each of seven types (A to G) of classification patterns as shown in FIG. Generate attribute classification information for types.

When such a plurality of attribute-based classification information are stored in the database, for example, the acquisition unit 11 acquires evaluation data corresponding to the user's acceleration pulse wave and preliminary evaluation data. The generator 12 then refers to the preliminary evaluation data and selects the first classification information. After that, the generation unit 12 refers to the first classification information and generates an evaluation result for the evaluation data. Therefore, it is possible to select the most suitable classification information for the user from each attribute classification information.

For example, when the data corresponding to the velocity pulse wave of the subject is used as the input data for the learning data, the input data is input for each of the two types of classification patterns (group 1 and group 2) as shown in FIG. You may prepare and generate|occur|produce two types of attribute classification information.

Here, the data corresponding to the accelerated pulse wave shown in FIG. 7 is easy to classify in detail based on the characteristics, but there is a concern that accuracy may be lowered due to erroneous detection of the peak when calculating the blood sugar level. In addition, the data corresponding to the velocity pulse wave shown in FIG. 8 is more difficult to classify in detail based on the characteristics than the data corresponding to the acceleration pulse wave, but since there are few false detections of peaks, etc., it is possible to increase the blood sugar level. It can be calculated with precision.

Based on the above, a plurality of attribute classification information includes, for example, data corresponding to an acceleration pulse wave as shown in FIG. 7 as a classification pattern used to select specific classification information, and when generating attribute classification information Data corresponding to a velocity pulse wave may be used as the learning data.

In this case, as the acquisition step S110, for example, the acquisition unit 11 acquires data corresponding to the velocity pulse wave as evaluation data from sensor data based on the user's pulse wave. Further, the acquisition unit 11 acquires data corresponding to the acceleration pulse wave from the sensor data as preliminary evaluation data.

Next, as a selection step S121, for example, the generation unit 12 refers to the preliminary evaluation data, and among a plurality of classification patterns including data corresponding to the acceleration pulse wave, the classification pattern (first classification) most similar to the preliminary evaluation data. pattern), and select the first classification information linked to the first classification pattern. Then, as an attribute-based generation step S122, the generation unit 12 refers to the first classification information and generates an evaluation result for the evaluation data. This makes it possible to further improve the evaluation accuracy.

For example, the generator 12 determines which pattern in FIG. 7 the preliminary evaluation data applies to, and identifies the first classification pattern. For example, if the inflection point b of the input preliminary evaluation data is smaller than the inflection point d and the inflection point c≧0.5, the pattern A is set as the first classification pattern. Thereby, the blood sugar level can be calculated with high accuracy by referring to the classification information suitable for the characteristics of the evaluation data.

According to the present embodiment, in addition to the effects of the above-described embodiments, the generating unit 12 refers to the preliminary evaluation data and selects the first classification information, and refers to the first classification information to select the evaluation data. an attribute-based generator for generating blood glucose level evaluation results. Therefore, it is possible to select the attribute classification information that is most suitable for the characteristics of the pulse wave, and then generate the blood glucose level evaluation result for the evaluation data. This makes it possible to further improve the evaluation accuracy.

Further, according to the present embodiment, the acquisition unit 11 acquires data corresponding to velocity pulse waves based on pulse waves as evaluation data. Further, the acquiring unit 11 acquires data corresponding to an accelerated pulse wave based on the pulse wave as preliminary evaluation data. Therefore, the attribute classification information can be selected using the acceleration pulse wave, which makes it easier to classify the characteristics of the pulse wave than the velocity pulse wave. In addition, the blood sugar level evaluation result can be generated using the velocity pulse wave, which makes it easier to calculate the blood sugar level than the accelerated pulse wave. This makes it possible to further improve the evaluation accuracy.

While embodiments of the invention have been described, the embodiments have been presented by way of example and are not intended to limit the scope of the invention. Such novel embodiments can be implemented in various other forms, and various omissions, replacements, and modifications can be made without departing from the scope of the invention. Embodiments of the present invention may be formed by combining a plurality of embodiments among, for example, the first embodiment to the eighth embodiment. This embodiment and its modifications are included in the scope and gist of the invention, and are included in the scope of the invention described in the claims and its equivalents.

1: biological information computing device 3: communication network 4: server 5: sensor 6: detection unit 10: housing 11: acquisition unit 12: generation unit 13: evaluation unit 14: output unit 15: storage unit 16: learning unit 50: Acquisition unit 51: communication I/F
52: memory 53: instruction unit 54: internal bus 55: wristband 100: biological information calculation system 101: CPU
102: ROM
103: RAM
104: Storage unit 105: I/F
106: I/F
107: I/F
108: Input unit 109: Display unit 110: Internal bus S110: Acquisition step S120: Generation step S121: Selection step S122: Attribute-based generation step S130: Evaluation step S131: Evaluation selection step S132: Attribute-based evaluation step S140: Output step S150: Blood sugar trend step

A biological information calculation system according to a first aspect of the invention is a biological information calculation system that generates an evaluation result regarding a user's health condition, and includes any of a volume pulse wave, a velocity pulse wave, and an acceleration pulse wave based on the user's pulse wave. Acquisition means for acquiring evaluation data corresponding to , input data based on the learning pulse wave acquired in advance, and a pair of reference data including a blood glucose level linked to the input data as learning data, a plurality of the learning a database storing classification information generated using the data for the above; generating means for generating a blood glucose level evaluation result including the blood glucose level for the evaluation data by referring to the database; and the evaluation data and the blood glucose level and evaluation means for generating the evaluation result based on the evaluation result.

A biological information calculation system according to a first aspect of the invention is a biological information calculation system that generates an evaluation result regarding a user's health condition, and includes any of a volume pulse wave, a velocity pulse wave, and an acceleration pulse wave based on the user's pulse wave. and a pair of input data of the same type as the evaluation data based on the learning pulse wave acquired in advance and reference data including the blood glucose level linked to the input data as learning data a database that stores classification information generated using a plurality of the learning data; generating means that refers to the database and generates a blood glucose level evaluation result including the blood glucose level for the evaluation data; and evaluation means for generating the evaluation result based on the evaluation data and the blood glucose level evaluation result.

Claims (9)

A biological information computing system that generates an evaluation result regarding a user's health condition,
Acquisition means for acquiring evaluation data based on the user's pulse wave;
A pair of input data based on learning pulse waves acquired in advance and a pair of reference data including a blood glucose level linked to the input data are used as learning data, and classification information generated using a plurality of the learning data is stored. and
generating means for referring to the database and generating a blood sugar level evaluation result including the blood sugar level for the evaluation data;
and evaluation means for generating the evaluation result based on the evaluation data and the blood glucose level evaluation result. The evaluation means are
acquiring a blood sugar trend related to the tendency of the user's blood sugar level from time-series changes in a plurality of the blood sugar level evaluation results generated from the different evaluation data;
2. The biological information computing system according to claim 1, further comprising: generating said evaluation result from said blood glucose trend and said evaluation data. 3. The biological information computing system according to claim 1, wherein the evaluation means selects a function for generating the evaluation result from the blood glucose level evaluation result based on the characteristics of the evaluation data. . The evaluation means are
Acquiring additional information indicating characteristics of the user;
The biological information computing system according to any one of claims 1 to 3, further comprising: generating said evaluation result based on said evaluation data, said blood glucose level evaluation result and said additional information. The input data includes learning additional information indicating user characteristics associated with the learning pulse wave,
The generating means is
Acquiring additional information indicating characteristics of the user;
The biometric information computing system according to any one of claims 1 to 3, further comprising referring to the database and generating the blood glucose level evaluation result for the evaluation data and the additional information. The generating means is
Acquiring additional information indicating characteristics of the user;
selecting first classification information from the classification information based on the additional information;
The biological information computing system according to any one of claims 1 to 3, further comprising referring to the first classification information and generating the blood glucose level evaluation result for the evaluation data. The acquiring means comprises acquiring additional data indicating characteristics different from the evaluation data based on the pulse wave,
5. The biological information computing system according to claim 4, wherein said generating means includes generating said additional information based on said additional data. The obtaining means includes obtaining preliminary evaluation data different from the evaluation data based on the pulse wave,
The classification information includes a plurality of attribute-based classification information calculated using different learning data,
The generating means is
selecting means for selecting first classification information from among the plurality of attribute-based classification information with reference to the preliminary evaluation data;
2. The biometric information computing system according to claim 1, further comprising an attribute-based generation unit that refers to the first classification information and generates the blood glucose level evaluation result for the evaluation data. The acquisition means is
Acquiring data corresponding to the velocity pulse wave based on the pulse wave as the evaluation data,
9. The biological information computing system according to claim 8, further comprising acquiring data corresponding to an accelerated pulse wave based on said pulse wave as said preliminary evaluation data.

Images