Biological information computing system, server, and data structure

Abstract

[Problem] To provide a biological information computing system capable of improving accuracy in evaluation of biological information. [Solution] A biological information computing system evaluates biological information of a user, the biological information computing system having: an acquisition unit that acquires first evaluation data and second evaluation data based on a pulse wave of the user; a database; and a generation unit that generates a first evaluation result including first biological information corresponding to the first evaluation data with reference to the database. Classification information is stored in the database, the classification information being generated using a plurality of pairs of input data and reference data as learning data, wherein the input data is data based on a pulse wave acquired in advance for learning, and the reference data is data including biological information associated with the input data.

Description

Technical Field

[0001] This invention relates to biological information processing systems, servers, and data structures. Background Technology

[0002] Previously, methods for evaluating users' biological information, such as those in Patent Document 1, have been proposed.

[0003] Patent document 1 discloses a biological information estimation device and method that measures a user's acceleration pulse wave and extracts the user's blood glucose information in a non-invasive manner without using spectrophotometry, based on the correlation between the blood glucose value measured by an invasive measurement method according to the waveform information of the measured acceleration pulse wave and the simultaneously measured acceleration pulse wave.

[0004] Existing technical documents

[0005] Patent documents

[0006] Patent Document 1: Japanese Patent No. 6544751 Summary of the Invention

[0007] The problem that the invention aims to solve

[0008] Here, the goal is to improve the accuracy of evaluating users' biometric information.

[0009] Therefore, the present invention is proposed in view of the above-mentioned problems, and its purpose is to provide a biological information computing system, server and data structure that can improve the accuracy of evaluating biological information.

[0010] Methods for solving problems

[0011] The first invention discloses a bio-information computing system for evaluating a user's bio-information. The bio-information computing system is characterized by comprising: an acquisition unit that acquires first evaluation data and second evaluation data based on the user's pulse wave; a database storing classification information generated by using multiple learning data pairs as learning data, wherein the input data is based on pre-acquired learning pulse wave data, and the reference data is data containing bio-information associated with the input data; and a generation unit that, with reference to the database, generates a first evaluation result containing first bio-information corresponding to the first evaluation data.

[0012] The biological information processing system of the second invention is characterized in that, in the first invention, the generation unit generates a second evaluation result containing a second type of biological information of a different kind from the first biological information, which corresponds to the second evaluation data.

[0013] The biological information processing system of the third invention is characterized in that, in the second invention, the biological information processing system further has a storage unit, which stores the first evaluation result and the second evaluation result.

[0014] The biological information processing system of the fourth invention is characterized in that, in the second invention, the acquisition unit performs different types of processing on a pulse wave data corresponding to either the velocity pulse wave or the acceleration pulse wave based on the user's pulse wave, thereby obtaining the first evaluation data and the second evaluation data.

[0015] The biological information computing system of the fifth invention is characterized in that, in the second invention, the classification information includes first classification information and second classification information generated using different types of learning data, the generation unit refers to the first classification information to generate the first evaluation result corresponding to the first evaluation data, and the generation unit refers to the second classification information to generate the second evaluation result corresponding to the second evaluation data.

[0016] The biological information computing system of the sixth invention is characterized in that, in the second invention, the biological information computing system has a comprehensive evaluation unit, which obtains additional information representing the characteristics of the user, and generates a comprehensive evaluation result that comprehensively evaluates the characteristics of the user based on the first evaluation result, the second evaluation result, and the additional information.

[0017] The biological information processing system of the seventh invention is characterized in that, in the first invention, the classification information includes multiple attribute-based classification information calculated using different learning data, and the generation unit includes: a selection unit that selects a first classification information from the multiple attribute-based classification information with reference to the second evaluation data; and an attribute-based generation unit that generates a first evaluation result corresponding to the first evaluation data with reference to the first classification information.

[0018] The biological information processing system of the eighth invention is characterized in that, in the seventh invention, the acquisition unit acquires data equivalent to the velocity pulse wave based on the pulse wave as the first evaluation data, and the acquisition unit acquires data equivalent to the acceleration pulse wave based on the pulse wave as the second evaluation data.

[0019] The biological information computing system of the ninth invention is characterized in that, in the first invention, the generation unit selects a first classification information from the classification information based on the characteristics of the pulse wave, and the generation unit generates the first evaluation result corresponding to the first evaluation data by referring to the first classification information.

[0020] The biological information computing system of the tenth invention is characterized in that, in the third invention, the biological information computing system has a computing unit, which generates a comprehensive evaluation result that comprehensively evaluates the characteristics of the user based on the first evaluation result and the second evaluation result stored by the storage unit.

[0021] The biological information computing system of the 11th invention is characterized in that, in the 3rd invention, the storage unit obtains the judgment result of the user's judgment on the content of the first evaluation result and the second evaluation result, and the storage unit associates and stores the judgment result, the first evaluation result and the second evaluation result respectively.

[0022] The biological information processing system of the 12th invention is characterized in that, in the 11th invention, the biological information processing system has an update unit, which updates the classification information based on the judgment result, the first evaluation result and the second evaluation result stored in the storage unit.

[0023] The biological information processing system of the 13th invention is characterized in that, in the 1st invention, the biological information processing system has: a server that stores the database and generates the first evaluation result through the generation unit; and a biological information processing device that receives the first evaluation result from the server and displays it.

[0024] The server of the 14th invention is characterized in that it stores the first evaluation result as described in claim 1.

[0025] The data structure of the 15th invention is a data structure used in a computer having a display unit, a control unit, and a storage unit, and is stored in the storage unit. The data structure is characterized in that it includes the first evaluation result and the second evaluation result generated by the biological information computing system of the 2nd invention, and uses the first evaluation result and the second evaluation result when the control unit generates a comprehensive evaluation result that comprehensively evaluates the characteristics of the user.

[0026] Invention Effects

[0027] According to the first to thirteenth inventions, the acquisition unit acquires first evaluation data and second evaluation data based on a user's pulse wave. Furthermore, the generation unit generates a first evaluation result containing first biological information corresponding to the first evaluation data. That is, multiple evaluation data are acquired based on the pulse wave of one user. Therefore, when acquiring each evaluation data, deviations caused by the pulse wave measurement conditions can be eliminated. This improves the accuracy of evaluating biological information.

[0028] According to the 1st to 13th inventions, the generation unit generates a first evaluation result with reference to a database. Furthermore, the database stores classification information generated using multiple learning data. Therefore, when generating the evaluation result, a quantitative evaluation result can be generated based on the relationship between the characteristics of past pulse waves and biological information. This helps to suppress biases arising from subjective evaluations by users, etc.

[0029] Specifically, according to the 13th invention, the server generates the first evaluation result through the generation unit. Furthermore, the biological information processing device receives the first evaluation result from the server and displays it. Therefore, the workload of generating evaluation results can be reduced for the biological information processing device. This improves the convenience of the biological information processing device. Furthermore, it is not necessary to store a database in the biological information processing device. This significantly reduces the data storage capacity of the biological information processing device. Moreover, by storing the database on the server, evaluation results generated with reference to one classification information can be output to multiple biological information processing devices. This reduces the enormous time and cost associated with updating the database for each biological information processing device, which is required for maintenance such as database updates.

[0030] According to the 14th invention, the server is able to store the first evaluation result, which achieves improved accuracy when evaluating biological information.

[0031] According to the 15th invention, the data structure includes a first evaluation result and a second evaluation result that improve the accuracy of evaluating biological information, and the data structure can be used when generating a comprehensive evaluation result. Attached Figure Description

[0032] Figure 1 This is a schematic diagram illustrating an example of the biological information processing system in the first embodiment.

[0033] Figure 2 This is a schematic diagram illustrating an example of the operation of the biological information processing system in the first embodiment.

[0034] Figure 3 (a) is a schematic diagram illustrating an example of classification information. Figure 3(b) is a schematic diagram illustrating another example of categorized information.

[0035] Figure 4 (a)~ Figure 4 (d) is a schematic diagram illustrating an example of processing sensor data.

[0036] Figure 5 (a) is a schematic diagram illustrating an example of the structure of a biological information processing device. Figure 5 (b) is a schematic diagram illustrating an example of the function of a biological information processing device.

[0037] Figure 6 (a) and Figure 6 (b) is a schematic diagram showing an example of a sensor.

[0038] Figure 7 This is a flowchart illustrating an example of the operation of the biological information processing system in the first embodiment.

[0039] Figure 8 This is a schematic diagram illustrating an example of the operation of the biological information processing system in the second embodiment.

[0040] Figure 9 This is a schematic diagram illustrating a modified example of the operation of the biological information processing system in the second embodiment.

[0041] Figure 10 This is a schematic diagram illustrating an example of the operation of the biological information processing system in the third embodiment.

[0042] Figure 11 This is a schematic diagram illustrating an example of the operation of the biological information processing system in the fourth embodiment.

[0043] Figure 12 This is a schematic diagram showing a classification example of data equivalent to acceleration pulse waves.

[0044] Figure 13 This is a schematic diagram showing a classification example of data equivalent to a velocity pulse wave.

[0045] Figure 14 This is a schematic diagram illustrating a first variation of the operation of the biological information processing system in the fourth embodiment.

[0046] Figure 15 This is a schematic diagram illustrating a second variation of the operation of the biological information processing system in the fourth embodiment.

[0047] Figure 16 This is a schematic diagram illustrating a third variation of the operation of the biological information processing system in the fourth embodiment.

[0048] Figure 17 This is a schematic diagram illustrating an example of the operation of the biological information processing system in the fifth embodiment.

[0049] Figure 18 This is a schematic diagram illustrating an example of the operation of the biological information processing system in the sixth embodiment.

[0050] Figure 19 This is a diagram showing the sequence of biological information processing procedures used to implement the electronic device in the 7th embodiment.

[0051] Figure 20 This is a schematic diagram showing an example of the first extraction unit and the first data acquisition unit.

[0052] Figure 21 This is a schematic diagram showing an example of the second extraction unit and the second data acquisition unit.

[0053] Figure 22 This is a flowchart illustrating an example of the operation of the biological information processing system in the seventh embodiment.

[0054] Figure 23 This is a diagram showing the sequence of biological information processing procedures used to implement the electronic device in the eighth embodiment.

[0055] Figure 24 This is a schematic diagram illustrating an example of a categorized data extraction unit.

[0056] Figure 25 This is a schematic diagram illustrating an example of the data extraction unit for evaluation.

[0057] Figure 26 This is a flowchart illustrating an example of the operation of the biological information processing system in the 8th embodiment. Detailed Implementation

[0058] Hereinafter, an example of a biological information processing system, server, and data structure according to an embodiment of the present invention will be described with reference to the accompanying drawings.

[0059] (First Embodiment: Biological Information Processing System 100)

[0060] Figure 1 This is a schematic diagram illustrating an example of the biological information processing system 100 in the first embodiment.

[0061] The bio-information processing system 100 is used to evaluate a user's bio-information. The system acquires multiple evaluation data points based on the user's pulse wave and generates an evaluation result containing bio-information corresponding to the evaluation data. That is, the multiple evaluation data points are obtained based on the pulse wave of a single user. Therefore, when acquiring each evaluation data point, biases caused by the pulse wave measurement conditions can be eliminated. For example, the processing conditions for other evaluation data can be determined based on the characteristics of one evaluation data point. Furthermore, different evaluation results can be generated separately for each of the multiple evaluation data points. Regardless of the method, the system uses the pulse wave of a single user, thus suppressing the influence of measurement conditions. Therefore, the accuracy of evaluating bio-information can be improved.

[0062] Specifically, the bio-information computing system 100 in this embodiment generates multiple evaluation results, each containing different types of bio-information, based on multiple evaluation data from the user's pulse wave. That is, it can generate evaluation results containing various types of bio-information based on a single pulse wave measured from the user. Furthermore, "bio-information" refers, for example, to characteristics of the blood and body that can be estimated based on the pulse wave.

[0063] As biological information, in addition to characteristics such as blood carbon dioxide levels, other indicators include blood glucose levels, blood pressure, oxygen saturation, lactate levels, pulse rate, respiratory rate, stress levels, vascular age, and the degree of diabetes. Furthermore, "characteristics of blood carbon dioxide" indicates the level of carbon dioxide in the user's blood. As characteristics of blood carbon dioxide, in addition to using the partial pressure of carbon dioxide (PaCO2), other indicators include the dissolved concentration of carbon dioxide in the blood and the amount of bicarbonate / bicarbonate (HCO3) in the blood. - The concentration of ) can also be determined based on the situation, taking into account the pH value of the blood.

[0064] For example, Figure 1 As shown, the biological information processing system 100 includes a biological information processing device 1, and may also include at least one of a sensor 5 and a server 4. The biological information processing device 1 is connected to the sensor 5 and the server 4, for example, via a communication network 3.

[0065] In the biological information processing system 100, for example, Figure 2As shown in (a), the bio-information processing device 1 acquires sensor data generated by the sensor 5, etc. The sensor data represents data containing characteristics of a user's pulse wave measured within a specific period. Then, the bio-information processing device 1 performs preprocessing such as filtering on the acquired sensor data to obtain multiple evaluation data (e.g., first evaluation data and second evaluation data). That is, when acquiring each evaluation data, the bio-information processing device 1 performs different types of preprocessing on each sensor data.

[0066] The organism information processing device 1 refers to a database and calculates different types of organism information (e.g., first organism information and second organism information) for each evaluation data. Then, the organism information processing device 1 generates multiple evaluation results (e.g., first evaluation result and second evaluation result) each containing different types of organism information. That is, multiple organism information are calculated based on the pulse wave of a single user. Therefore, when calculating each organism information, deviations caused by the pulse wave measurement conditions can be eliminated. This improves the accuracy of evaluating organism information.

[0067] Here, the organism information processing device 1 refers to a database when calculating the organism information corresponding to each evaluation data. The database stores classification information generated using multiple learning data.

[0068] For example, Figure 3 As shown in (a), the classification information is generated using multiple learning data sets by using a pair of input data based on previously acquired pulse waves and reference data containing biological information associated with the input data. Therefore, when calculating biological information, quantitative evaluation results can be generated based on the relationship between the characteristics of previously proven pulse waves and biological information. This helps to suppress biases arising from subjective evaluations by users, etc.

[0069] In addition, for example, Figure 3 As shown in (b), the classification information can also include multiple classification information generated using different types of learning data (e.g., first classification information and second classification information). In this case, evaluation results can be generated by referring to the best classification information according to the type of each evaluation data.

[0070] The biological information processing device 1 stores the generated evaluation results on the server 4, etc. This allows for easy secondary use of each evaluation result.

[0071] The biological information processing device 1 outputs the generated evaluation results to a display, for example. This allows the user to access the multiple evaluation results generated.

[0072] Alternatively, in the biological information processing system 100, multiple evaluation data can be obtained from sensors such as sensor 5. In this case, the preprocessing of obtaining multiple evaluation data from sensor data is performed by sensor 5 or the like.

[0073] <Sensor Data>

[0074] Sensor data includes data representing characteristics of the user's pulse wave, and may also include data representing characteristics other than the pulse wave (noise). Sensor data represents amplitude data corresponding to the measurement time. By implementing filtering processing appropriate to the application and the conditions under which the sensor data is generated, data equivalent to acceleration pulse waves, velocity pulse waves, etc., can be obtained from the sensor data.

[0075] Sensor data can be generated by known sensors such as strain sensors, gyroscope sensors, photoplethysmography (PPG) sensors, and pressure sensors. In addition to digital signals, sensor data can also be analog signals. Furthermore, the measurement time for generating sensor data can be arbitrarily set, for example, a measurement time corresponding to 1 to 20 cycles of a pulse wave, depending on the sensor data processing method, data communication method, and other conditions.

[0076] <Evaluation Data>

[0077] Evaluation data represents data used to calculate information about a living organism. For example, evaluation data may represent data equivalent to an acceleration pulse wave based on a user's pulse wave, representing the amplitude corresponding to a specific period (e.g., one period).

[0078] The evaluation data is obtained by processing (preprocessing) the sensor data using a biological information processing device 1, etc. For example, Figure 4 (a)~ Figure 4 As shown in (d), evaluation data can be obtained by performing multiple processes on the sensor data. Details of each process will be explained later.

[0079] <Database>

[0080] Databases are primarily used when calculating and evaluating the results corresponding to the data. In addition to storing one or more classification entries, a database can also store multiple training datasets used to generate classification information.

[0081] Categorical information, for example, is a function representing the correlation between previously obtained evaluation data (input data) and reference data containing biological information. Categorical information can also represent a calibration model generated by using regression analysis, with input data as explanatory variables and reference data as the target variable, and based on the analysis results. Regarding categorical information, in addition to periodically updating the calibration model, it can also be generated based on user attributes such as gender, age, and exercise content.

[0082] Methods of regression analysis used in generating classification information include, for example, PLS (Partial Least Squares) regression analysis and SIMCA (Soft Independent Modeling of Class Analogy) regression analysis, which uses principal component analysis to obtain principal component models for each category.

[0083] Classification information can also include, for example, a learned model generated by machine learning using multiple training datasets. This learned model can represent neural network models such as CNNs (Convolutional Neural Networks), as well as SVMs (Support Vector Machines). Furthermore, deep learning can be used as a machine learning method.

[0084] The input data uses the same type of data as the evaluation data, such as past evaluation data that clearly represents the corresponding biological information. For example, the subject wears sensor 5, etc., and sensor data (learning sensor data) representing the characteristics of the learning pulse wave is generated. Then, by processing the learning sensor data, the input data can be obtained. In addition to obtaining input data from the user of the biological information processing system 100, input data can also be obtained from users different from that user. That is, the subject mentioned above can be not only the user of the biological information processing system 100, but also users other than that user, or multiple users, either specific or non-specific.

[0085] The input data is preferably obtained in accordance with the same type of sensor (e.g., sensor 5) used when obtaining the evaluation data, the same conditions for generating the sensor data, and the same conditions for processing the sensor data. For example, by unifying the above three aspects, the accuracy of generating biological information can be significantly improved.

[0086] The reference data includes biological information of the subject measured using a measuring device. For example, when generating learning sensor data by having the subject wear sensor 5, reference data associated with the input data can be obtained by measuring biological information such as the characteristics of carbon dioxide in the subject's blood. In this case, it is preferable that the timing of measuring the biological information and the timing of generating the learning sensor data are simultaneous, but for example, it is also possible that they are measured approximately 1 to 10 minutes apart.

[0087] The reference data are measured using known measuring devices. For example, when measuring the characteristic concentration of carbon dioxide in the blood, a transcutaneous blood gas monitor such as the TCM5 (Radiometer Basel) is used as the measuring device. Similarly, when measuring the amount of lactate in the blood, a known measuring device such as the Lactate Pro 2 (ARKRAY) is used. And when measuring oxygen saturation, a known measuring device such as the PULSOX-Neo (KONICAMINOLTA) is used.

[0088] <Biological Information>

[0089] The organism information calculated in the organism information processing system 100 is calculated as data of the same type as the reference data. The organism information is calculated as data that is the same as or similar to the reference data, based on reference classification information. In the organism information processing system 100, for example, multiple evaluation data are acquired along any time series, and multiple organism information corresponding to each evaluation data is generated. Furthermore, in the organism information processing system 100, for example, multiple evaluation data can be acquired at any time, and multiple organism information corresponding to each evaluation data can be calculated.

[0090] <Evaluation Results>

[0091] The evaluation result includes the user's biometric information calculated by the biometric information processing device 1. In addition to biometric information, the evaluation result may also include comparisons with pre-set thresholds. The evaluation result may also include values ​​derived from multiple biometric data over a time series. For example, the evaluation result may include information such as "healthy," "needs exercise," or "undergoing anaerobic exercise," indicating the user's health status, suggestions for maintaining health, and assessments corresponding to exercise capacity. Thus, it is easy to grasp the temporal changes of multiple biometric data. For example, by outputting the evaluation result, the user's biometric information can be grasped.

[0092] <Biological Information Processing Device 1>

[0093] The bio-information processing device 1 may refer to electronic devices such as personal computers (PCs), mobile phones, smartphones, tablet terminals, and wearable terminals, or electronic devices capable of communicating via communication network 3 based on user operations. Furthermore, the bio-information processing device 1 may also have a built-in sensor 5. An example of using a PC as the bio-information processing device 1 will be described below.

[0094] Figure 5 (a) is a schematic diagram showing an example of the structure of the biological information processing device 1. Figure 5 (b) is a schematic diagram illustrating an example of the function of the biological information processing device 1.

[0095] For example, Figure 5 As shown in (a), the biological information processing 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 / F units 105 to 107. Each of the structures 101 to 107 is connected via an internal bus 110.

[0096] CPU 101 controls the entire biological information processing device 1. ROM 102 stores the operation code of CPU 101. RAM 103 is the working area used when CPU 101 is operating. Storage unit 104 stores various information such as databases and evaluation data. As storage unit 104, in addition to HDD (Hard Disk Drive), SSD (Solid State Drive) or other data storage devices may be used. Alternatively, biological information processing device 1 may also have a GPU (Graphics Processing Unit) (not shown).

[0097] I / F105 is an interface for transmitting and receiving various information with the server 4, sensor 5, etc., as needed via the communication network 3. I / F106 is an interface for transmitting and receiving information with the input unit 108. As the input unit 108, for example, a keyboard is used, and users of the bio-information processing device 1 input various information or control commands of the bio-information processing device 1 via the input unit 108. I / F107 is an interface for transmitting and receiving various information with the display unit 109. The display unit 109 displays various information or evaluation results stored in the storage unit 104. As the display unit 109, a display is used, for example, a touch panel type, and it is integrated with the input unit 108.

[0098] Figure 5 (b) is a schematic diagram illustrating an example of the function of the biological information processing device 1. The biological information processing device 1 includes an acquisition unit 11, a generation unit 12, an evaluation unit 13, and an output unit 14; for example, it may also include a learning unit 15. Furthermore, Figure 5 The functions shown in (b) are implemented by CPU 101 using RAM 103 as a working area to execute programs stored in storage unit 104, etc.

[0099] <Acquisition Section 11>

[0100] The acquisition unit 11 acquires multiple evaluation data based on the user's pulse wave. For example, after acquiring sensor data from sensor 5, the acquisition unit 11 processes the sensor data to obtain evaluation data. The acquisition unit 11 performs different types of processing to acquire multiple evaluation data for a single sensor data set.

[0101] For example, Figure 4 As shown in (b), the acquisition unit 11 performs filtering processing (filtering) on ​​the acquired sensor data. In the filtering process, for example, a bandpass filter of 0.5 to 5.0 Hz is used. Thus, the acquisition unit 11 extracts data corresponding to the user's pulse wave (pulse wave data). The pulse wave data represents, for example, data corresponding to a velocity pulse wave. Furthermore, the pulse wave data can also represent, for example, data corresponding to an acceleration pulse wave or a volume pulse wave, and can be arbitrarily set according to the type and application of the sensor. In addition, the filtering range of the bandpass filter can be arbitrarily set according to the application.

[0102] The acquisition unit 11 performs differential processing on pulse wave data, for example. For example, when differential processing is performed on pulse wave data that corresponds to velocity pulse waves, the acquisition unit 11 acquires data (differential data) that corresponds to acceleration pulse waves. Furthermore, in the differential processing, in addition to first-order differential processing, second-order differential processing may also be performed.

[0103] The acquisition unit 11 performs segmentation processing on differential data, for example. In the segmentation processing, differential data corresponding to multiple cycles of acceleration pulse waves is segmented into data corresponding to each cycle of acceleration pulse waves (segmented data). Therefore, the acquisition unit 11 can acquire multiple segmented data by performing segmentation processing on one differential data. Furthermore, in the segmentation processing, the differential data can be segmented according to any period (e.g., a positive multiple of the period) depending on the application.

[0104] For example, in the segmentation process, there may be cases where the amount of data in each segment differs. In such cases, the acquisition unit 11 can also determine the segment with the least amount of data and perform data reduction (trimming) on ​​the other segments. As a result, the amount of data in each segment can be standardized, making it easier to compare the data in each segment.

[0105] In addition to the above, the values ​​corresponding to the time axis in the segmented data can also be normalized. Normalization could involve setting the minimum value corresponding to the time axis to 0 and the maximum value to 1. This makes comparing the data in each segment easier.

[0106] The acquisition unit 11 may, for example, calculate the average of multiple segmented data that have undergone data reduction or normalization, and use it as the segmented data.

[0107] The acquisition unit 11 performs normalization processing on the segmented data. In the normalization process, normalized data (normalized data) is generated using values ​​corresponding to amplitude as objects. For example, the normalization process may involve setting the minimum amplitude value to 0 and the maximum amplitude value to 1. The acquisition unit 11 acquires the normalized data as evaluation data (e.g., evaluation data A). In this case, as evaluation data A, data corresponding to the user's acceleration pulse wave is obtained.

[0108] In addition to performing the aforementioned processes sequentially, the acquisition unit 11 may, for example, Figure 4 As shown in (b), differentiation may not be performed. In this case, data equivalent to the user's velocity pulse wave is obtained as evaluation data (e.g., evaluation data B).

[0109] Alternatively, the acquisition unit 11 may perform only a portion of the processes described above. In this case, the acquisition unit 11 may acquire any data from pulse wave data, differential data, segmented data, trimmed segmented data, and segmented data obtained by normalizing values ​​corresponding to the time axis as evaluation data, and can be arbitrarily set according to the purpose.

[0110] For example, Figure 4 As shown in (c), the acquisition unit 11 can also acquire evaluation data C that is suitable for calculating pulse count as information about a living organism.

[0111] In this case, the acquisition unit 11 performs the aforementioned filtering process 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 processing, multiple peaks (maximum amplitudes) contained in the pulse wave data are detected, and the sampling order (equivalent to the time from the start of the measurement) is determined. Thus, the acquisition unit 11 acquires the peak position data contained in the pulse wave data.

[0112] Then, the acquisition unit 11 performs peak interval averaging calculation processing on the peak position data. In the peak interval averaging calculation processing, the interval between the peaks contained in the peak position data (the difference in the sampling order of adjacent peaks) is calculated, for example, the average value of the peak intervals is calculated. Then, the acquisition unit 11 divides the peak interval or the average value of the peak intervals by the sampling rate of the sensor data to obtain data representing the peak intervals equivalent to the number of seconds as evaluation data (e.g., evaluation data C).

[0113] For example, Figure 4 As shown in (d), the acquisition unit 11 can also acquire evaluation data D that is suitable for calculating the respiratory rate as information about the organism.

[0114] In this case, the acquisition unit 11 performs the aforementioned filtering process on the sensor data to extract pulse wave data. Then, the acquisition unit 11 performs Fourier transform processing on the pulse wave data. In the Fourier transform processing, for example, the pulse wave data representing the amplitude relative to the sampling time is transformed into frequency data representing the intensity relative to the frequency. Thus, the acquisition unit 11 acquires the frequency data corresponding to the pulse wave data.

[0115] Next, the acquisition unit 11 performs maximum frequency detection processing on the frequency data. In the maximum frequency detection processing, the frequency with the maximum intensity between 0.15 and 0.35 Hz in the frequency data is determined. As a result, the acquisition unit 11 obtains the value of the determined frequency as evaluation data D.

[0116] <Generation Section 12>

[0117] The generation unit 12 generates evaluation results corresponding to the evaluation data by referring to a database. For example, the generation unit 12 calculates the organism information corresponding to the evaluation data by referring to classification information stored in the database, and generates the evaluation results accordingly. The generation unit 12 generates multiple evaluation results, each corresponding to one of the many different evaluation data sets.

[0118] The generation unit 12 can, for example, generate multiple evaluation data sets by referring to the same classification information. In this case, multiple evaluation data sets are obtained by using different types of preprocessing, so that even when using the same classification information, information about different types of organisms can be calculated separately. Therefore, it is not necessary to generate classification information for each evaluation data set, thus reducing the data volume of the database.

[0119] The generation unit 12 can, for example, calculate information on multiple organisms by referring to different classification information for multiple evaluation data. In this case, it is possible to generate each evaluation result by referring to the optimal classification information for each type of evaluation data.

[0120] The generation unit 12 may, for example, use a display format pre-stored in the storage unit 104, etc., to generate evaluation results that have been converted into a form that allows users to understand biological information.

[0121] <Output Section 13>

[0122] The output unit 13 outputs multiple evaluation results. In addition to outputting multiple evaluation results to the display unit 109, the output unit 13 can also output multiple evaluation results to, for example, the sensor 5.

[0123] <Storage Department 14>

[0124] The storage unit 14 retrieves various data, such as databases, stored in the storage unit 104 as needed. The storage unit 14 also stores various data acquired or generated by the structures 11-13 and 15 in the storage unit 104 as needed.

[0125] Storage unit 14 may also store sensor data associated with multiple evaluation results in storage unit 104. For example, storage unit 14 may also store multiple evaluation data generated using sensor data in storage unit 104.

[0126] <Study Department 15>

[0127] The learning unit 15 may generate classification information using multiple learning data sets, for example. The learning unit 15 may also acquire new learning data and update existing classification information.

[0128] <Communication Network 3>

[0129] Communication network 3 is a known internet or similar network that connects to the bio-information processing device 1, server 4, and sensor 5 via communication lines. When the bio-information processing system 100 is used within a limited area, communication network 3 can also be constructed using a LAN (Local Area Network). Furthermore, communication network 3 can also be constructed using a fiber optic communication network. Moreover, communication network 3 is not limited to wired communication networks; it can also be implemented via wireless communication networks, and can be arbitrarily configured according to the intended use.

[0130] <Server 4>

[0131] Server 4 stores and accumulates various information, such as evaluation results, sent via communication network 3. Based on requests from the biological information processing device 1, server 4 sends the accumulated information to the biological information processing device 1 via communication network 3.

[0132] Server 4 can, for example, connect to multiple biological information processing devices 1, obtain various information such as evaluation results from each biological information processing device 1, and store them together. Furthermore, server 4 may also possess at least some of the functions of the biological information processing device 1 described above. In addition, server 4 may store the database stored in the biological information processing device 1, etc.

[0133] <Sensor 5>

[0134] Sensor 5 generates sensor data. For example, Figure 6 As shown in (a), the sensor 5 has a detection unit 6. The sensor 5 is mounted at a position where the user's pulse wave can be detected via the detection unit 6, for example, fixed to a wristband 55.

[0135] The detection unit 6 uses a known detection device capable of detecting the user's pulse wave. For example, the detection unit 6 may use at least one of the following: a strain sensor such as a fiber Bragg grating (FBG) sensor, a gyroscope sensor, one or more electrodes for measuring pulse wave signals, a photoplethysmography (PPG) sensor, a pressure sensor, and an optical detection module. Multiple detection units 6 may also be configured, for example.

[0136] Furthermore, sensor 5 can also be embedded in clothing. In addition to humans, users wearing sensor 5 can also target pets such as dogs and cats, as well as livestock such as cows or pigs, and farmed animals such as fish.

[0137] For example, Figure 6 As shown in (b), the sensor 5 includes an acquisition unit 50, a communication I / F 51, a memory 52, and a command unit 53, and each of these components is connected via an internal bus 54.

[0138] The acquisition unit 50 measures the user's pulse wave via the detection unit 6 and generates sensor data. The acquisition unit 50 then sends the generated sensor data to, for example, a communication I / F 51 or a memory 52.

[0139] The communication I / F51 transmits various data, such as sensor data, to the biological information processing device 1 and the server 4 via the communication network 3. Additionally, the communication I / F51 is equipped with line control circuitry for connecting to the communication network 3, and signal conversion circuitry for data communication between the biological information processing device 1 and the server 4. The communication I / F51 performs conversion processing on various commands from the internal bus 54 and sends them to the communication network 3. Upon receiving data from the communication network 3, it performs prescribed conversion processing and sends it to the internal bus 54.

[0140] The memory 52 stores various data, such as sensor data, sent from the acquisition unit 50. For example, the memory 52 can send the stored sensor data and other data to the communication I / F 51 by receiving commands from other terminal devices connected via the communication network 3.

[0141] The command unit 53 includes operation buttons, a keyboard, and a processor such as a CPU for acquiring sensor data. When the command unit 53 receives a command to acquire sensor data, it notifies the acquisition unit 50. Upon receiving the notification, the acquisition unit 50 acquires the sensor data. Furthermore, for example... Figure 4 (a)~ Figure 4 As shown in (d), the command unit 53 can also perform processing to obtain evaluation data based on sensor data.

[0142] Here, as an example of obtaining sensor data, we will explain the case of using an FBG sensor.

[0143] The FBG sensor forms a diffraction grating structure at predetermined intervals within a single optical fiber. For example, the FBG sensor has the following characteristics: a sensor portion length of 10 mm, a wavelength resolution of ±0.1 pm, a wavelength range of 1550 ± 0.5 nm, an optical fiber diameter of 145 μm, and a core diameter of 10.5 μm. The FBG sensor can be used as the aforementioned detection unit 6, allowing it to perform measurements while in contact with the user's skin.

[0144] As a light source for the optical fiber, an ASE (Amplified Spontaneous Emission) source with a wavelength range of 1525–1570 nm is used, for example. The emitted light from the light source is incident on the FBG sensor via a circulator. The reflected light from the FBG sensor is guided by the circulator to a Mach-Zehnder interferometer, and the output light from the Mach-Zehnder interferometer is detected by a photodetector. A Mach-Zehnder interferometer is an instrument used to split a light beam into two optical paths with optical path difference using a beam splitter and then superimpose them into one using another beam splitter to generate interference light. To create the optical path difference, the length of one of the optical fibers can be increased, for example. The coherent light produces interference fringes based on the optical path difference; therefore, by measuring the pattern of the interference fringes, the strain change, i.e., the pulse wave, generated in the FBG sensor can be detected. The acquisition unit 50 generates sensor data based on the detected pulse wave. Thus, sensor data is acquired.

[0145] Furthermore, fiber optic sensor systems that detect pulse wave waveforms by detecting the strain of an FBG sensor include, in addition to a light source that directs the light beam onto the FBG sensor, a broadband ASE light source, a circulator, a Mach-Zehnder interferometer, a beam splitter, and other optical systems, a photodetector with a light-receiving sensor, and a resolution unit for resolving wavelength offset. In fiber optic sensor systems, the light source and frequency band can be selected according to the characteristics of the FBG sensor used, and various resolution methods, such as detection methods, can be employed.

[0146] Data Structures

[0147] For example, a data structure containing multiple evaluation results (e.g., a first evaluation result and a second evaluation result) generated by the aforementioned biological information processing system 100 is stored in the server 4 or the storage unit 104. This data structure is used in the aforementioned biological information processing device 1 (a computer having a display unit 109, a CPU 101 (control unit), and a storage unit 104). For example, when the generation unit 12, controlled by the CPU 101, generates a comprehensive evaluation result based on each evaluation result, the data structure containing multiple evaluation results is used. The comprehensive evaluation result will be explained later.

[0148] (First Embodiment: Operation of the Biological Information Processing System 100)

[0149] Next, an example of the operation of the biological information processing system 100 in this embodiment will be described. Figure 7 This is a flowchart illustrating an example of the operation of the biological information processing system 100 in this embodiment.

[0150] The biological information processing system 100 is executed, for example, by a biological information processing program installed in the biological information processing device 1. That is, the user can operate the biological information processing device 1 or the sensor 5, and obtain multiple evaluation results containing the user's biological information from the sensor data through the biological information processing program installed in the biological information processing device 1.

[0151] The biological information processing system 100 has an acquisition step S110, a generation step S120, and a storage step S140. For example, it may also have an output step S130.

[0152] <Obtaining step S110>

[0153] In step S110, multiple evaluation data are acquired based on the user's pulse wave. For example, the acquisition unit 50 of sensor 5 measures the user's pulse wave via the detection unit 6 and generates sensor data. The acquisition unit 50 transmits the sensor data to the biological information processing device 1 via communication I / F 51 and communication network 3. The acquisition unit 11 of the biological information processing device 1 receives the sensor data from sensor 5.

[0154] The acquisition unit 11, for example, performs sensor data acquisition. Figure 4 The processes shown in (a) and (b) of 4 are used to obtain the first evaluation data and the second evaluation data. The acquisition unit 11 stores each evaluation data in the storage unit 104 via the storage unit 14, for example. In addition, the frequency and other conditions of the acquisition unit 11 acquiring sensor data from the sensor 5 can be arbitrarily set according to the application. For example, the acquisition unit 11 acquires each evaluation data at a preset period.

[0155] <Generation Step S120>

[0156] Next, in generation step S120, multiple evaluation results containing organism information corresponding to each evaluation data are generated with reference to the database. For example, generation unit 12 calculates the blood carbon dioxide partial pressure value corresponding to the first evaluation data as the first organism information with reference to classification information, and calculates the blood glucose value corresponding to the second evaluation data as the second organism information. Generation unit 12 generates a first evaluation result containing the first organism information and a second evaluation result containing the second organism information.

[0157] The generation unit 12 generates, for example, a first evaluation result corresponding to the first evaluation data by referring to the first classification information, and a second evaluation result corresponding to the second evaluation data by referring to the second classification information. At this time, the first classification information and the second classification information are included in the classification information and are generated by using different types of learning data respectively.

[0158] The generation unit 12 saves each generated evaluation result to the storage unit 104 via the storage unit 14, for example. In addition, as for each evaluation result, in addition to showing specific values, the error range (e.g., "○○±2mmHg") can also be calculated.

[0159] For example, when calculating pulse rate as biological information, it can be used as evaluation data, such as... Figure 4 The generation unit 12 refers to the pulse count classification information included in the classification information for the evaluation data C shown in (c). The pulse count classification information represents, for example, a function of dividing 60 [seconds] by the peak interval. Therefore, the generation unit 12 can, for example, calculate the pulse count (=71 [bpm]) corresponding to the evaluation data C (peak interval = 0.85 [seconds]). Thus, the generation unit 12 can generate an evaluation result that includes biological information representing the pulse count.

[0160] For example, when calculating respiratory rate as information about a living organism, it can be used as evaluation data, such as... Figure 4 The evaluation data D shown in (d) is used by the generation unit 12 with reference to the respiratory rate classification information included in the classification information. The respiratory rate classification information is, for example, a function of multiplying the determined frequency by 60 [seconds]. Therefore, the generation unit 12 can, for example, calculate the respiratory rate (=13.5 [bpm]) corresponding to the evaluation data D (determined frequency = 0.225Hz). Thus, the generation unit 12 can generate an evaluation result containing information about the organism representing the respiratory rate.

[0161] <Output step S130>

[0162] Next, for example, in the output step S130, multiple evaluation results may also be output. For example, the output unit 13 outputs the first evaluation result and the second evaluation result to the display unit 109.

[0163] <Save step S140>

[0164] Next, in the saving step S140, the first evaluation result and the second evaluation result are saved. For example, the storage unit 14 saves the first evaluation result and the second evaluation result in the storage unit 104. For example, the output unit 13 may also output the first evaluation result and the second evaluation result to the server 4 via the communication network 3 for saving. For example, the saving step S140 may be performed before the output step S130.

[0165] Alternatively, in the saving step S140, sensor data associated with multiple evaluation results may also be saved, for example. In this case, the sensor data is used to generate multiple evaluation data representing the characteristics of the user's pulse wave.

[0166] Thus, the operation of the biological information processing system 100 is completed. Furthermore, the frequency and sequence of each step can be arbitrarily set according to the intended use.

[0167] In the biological information processing system 100, for example, in addition to implementing the aforementioned steps S110 and S120 in the biological information processing device 1, at least some of them may also be implemented in the server 4. In this case, the aforementioned steps S110 and S120 include processing by the server 4 for sending and receiving various information via the communication network 3. Furthermore, communication between the biological information processing device 1 and the server 4 can be implemented using known technologies.

[0168] For example, in step S110, server 4 may also acquire multiple evaluation data. In this case, the acquisition unit included in server 4 acquires multiple evaluation data sent from biological information processing device 1 via communication network 3.

[0169] The acquisition unit included in server 4 can, for example, acquire sensor data sent from biological information processing device 1 or sensor 5, and perform the aforementioned preprocessing to obtain multiple evaluation data. In this case, the workload of performing the aforementioned preprocessing can be reduced in biological information processing device 1.

[0170] For example, in generation step S120, server 4 can also generate multiple evaluation results corresponding to each evaluation data. In this case, the generation unit included in server 4 generates multiple evaluation results by referring to the database stored in server 4. In this case, the workload of performing the above-described process of generating multiple evaluation results can be reduced in the biological information processing device 1.

[0171] Furthermore, when the generation step S120 is performed in server 4, in the output step S130, for example, the output unit included in server 4 sends multiple evaluation results to the biological information processing device 1, etc., via communication network 3. In this case, the output unit 13 of biological information processing device 1 outputs the received multiple evaluation results to the display unit 109.

[0172] As described above, each step S110 to S140 in the biological information processing system 100 can be implemented by either the biological information processing device 1 or the server 4. In particular, by implementing the acquisition step S110 and the generation step S120 in the server 4, the load on the biological information processing device 1 can be reduced. Furthermore, the same applies to the embodiments described later, so the description will be omitted.

[0173] Here, when evaluating a user's biometric information, it is desirable to evaluate multiple biometric data points simultaneously. In existing technologies, measuring biometric data such as blood carbon dioxide saturation and hemoglobin relies on two volumetric pulse waves. Therefore, deviations in the measurement conditions of each volumetric pulse wave can significantly impact estimation accuracy. Furthermore, estimating multiple biometric data points using existing techniques requires obtaining multiple volumetric pulse waves corresponding to the number and type of biometric data. Therefore, the estimation accuracy due to deviations in each volumetric pulse wave may decrease proportionally to the number of volumetric pulse waves required to estimate multiple biometric data points.

[0174] In contrast, according to this embodiment, the acquisition unit 11 acquires first evaluation data and second evaluation data based on the user's pulse wave. Furthermore, the generation unit 12 generates a first evaluation result containing first organism information corresponding to the first evaluation data, and a second evaluation result containing second organism information of a different type than the first organism information corresponding to the second evaluation data. That is, the first and second organism information are calculated based on the pulse wave of one user. Therefore, when calculating each organism information, deviations caused by the pulse wave measurement conditions can be eliminated. This improves the accuracy of evaluating organism information.

[0175] Furthermore, according to this embodiment, the generation unit 12 generates a first evaluation result and a second evaluation result with reference to a database. In addition, the database stores classification information calculated using multiple learning data. Therefore, when generating each evaluation result, quantitative evaluation results can be generated based on the relationship between the characteristics of past pulse waves and biological information. This helps to suppress biases arising from subjective evaluations by users, etc.

[0176] Furthermore, according to this embodiment, the acquisition unit 11 performs different types of processing on a single pulse wave data point that corresponds to either the velocity pulse wave or the acceleration pulse wave based on the user's pulse wave, thereby acquiring first evaluation data and second evaluation data. Therefore, compared to the case using photoelectric volumetric pulse waves, evaluation can be performed using learning data and each evaluation data that suppresses the influence of noise data. This enables high-precision evaluation.

[0177] Furthermore, according to this embodiment, the classification information includes first classification information and second classification information generated using different types of learning data. Therefore, each evaluation result can be generated by referring to the optimal classification information based on the type of each evaluation data point. This allows for further improvement in the accuracy of evaluating biological information.

[0178] Furthermore, according to this embodiment, the classification information is obtained using a calibration model obtained through PLS regression analysis with input data as explanatory variables and reference data as target variables. Therefore, compared to calculating classification information using machine learning or similar methods, the amount of learning data can be significantly reduced, and the calibration model can be easily updated. This simplifies the construction and updating of the biological information processing system 100.

[0179] Furthermore, according to this embodiment, the first biological information represents blood glucose level, and the second biological information represents at least one of blood pressure, pulse rate, respiratory rate, blood carbon dioxide concentration, lactate level, and oxygen saturation. Therefore, compared to existing measurement methods, no invasive measurement method is required, and each piece of information can be easily obtained. This significantly reduces the burden on the user.

[0180] Furthermore, according to this embodiment, the storage unit 14 or the server 4 stores sensor data associated with the first evaluation result and the second evaluation result. Therefore, learning data can be easily prepared when classification information is updated or newly generated. As a result, the maintenance of the biological information processing system 100 can be easily achieved.

[0181] Furthermore, according to this embodiment, server 4 generates the first evaluation result and the second evaluation result through its included generation unit. The organism information processing device 1 receives the first evaluation result and the second evaluation result from server 4 and displays them. Therefore, the workload of generating each evaluation result can be reduced for the organism information processing device 1. This improves the convenience of the organism information processing device 1. Furthermore, it is not necessary to store a database in the organism information processing device 1. This significantly reduces the data storage capacity of the organism information processing device 1. Moreover, since the database is stored in server 4, evaluation results generated based on one classification information can be output to multiple organism information processing devices 1. This reduces the enormous time and cost associated with updating the database for each organism information processing device 1 during maintenance such as database updates.

[0182] Furthermore, according to this embodiment, server 4 is able to store the first evaluation result and the second evaluation result, which achieve improved accuracy when evaluating biological information.

[0183] Furthermore, according to this embodiment, the data structure includes a first evaluation result and a second evaluation result that improve the accuracy of evaluating biological information, and the first evaluation result and the second evaluation result can be used when generating a comprehensive evaluation result.

[0184] (Second Embodiment: Biological Information Processing System 100)

[0185] Next, an example of the biological information processing system 100 in the second embodiment will be described. The difference between the above-described embodiment and the second embodiment is that additional information is used in the second embodiment. Furthermore, descriptions of contents identical to those in the above-described embodiment will be omitted.

[0186] The organism information processing system 100 in this embodiment includes, for example, a comprehensive evaluation step S150. In the organism information processing system 100, for example, the comprehensive evaluation step S150 is performed after the generation step S120 described above, and the storage step S140 is performed after the comprehensive evaluation step S150.

[0187] In the comprehensive evaluation step S150, for example... Figure 8 As shown, additional information is obtained, and a comprehensive evaluation result is generated based on multiple evaluation results (e.g., the first evaluation result and the second evaluation result) and the additional information. The comprehensive evaluation step S150 can be performed, for example, by the comprehensive evaluation unit included in the generation unit 12.

[0188] Additional information represents user characteristics, such as information different from the biological information mentioned above. In this case, as additional information, in addition to using attribute information such as the user's gender and age, information that determines the user's health status, such as diagnostic results and exercise volume, can also be used. Regarding additional information, in addition to including information such as the user's medical history, lifestyle habits, medication information, degree of arteriosclerosis, and genetic information, it can also include information that may be related to the human body, such as the temperature and humidity of the surrounding environment and the acceleration measured by the accelerometer installed on sensor 5. Additional information is, for example, input by the user through input unit 108 and obtained by comprehensive acquisition unit.

[0189] The comprehensive evaluation results show the outcome of a holistic assessment of the user's characteristics. For example, the comprehensive evaluation results may show the results of corrections based on additional information applied to the biometric information included in each evaluation result. For instance, when using the user's age as additional information, a comparison between a pre-defined baseline value for each age group and each evaluation result is generated as the comprehensive evaluation result.

[0190] In addition to the above, as a comprehensive evaluation result, in addition to using strings to represent each user's characteristics such as "high blood sugar", "high blood pressure", "high exercise capacity" and "better at suppressing exercise", numerical values ​​such as the difference or deviation between the value and any benchmark value can also be used.

[0191] Furthermore, the comprehensive evaluation results may include, for example, estimated insurance information including the premium. Regarding the estimated insurance information, in addition to including values ​​representing the premiums estimated based on each evaluation result, it may also include strings indicating the type of insurance, etc. The estimated premium is calculated, for example, based on insurance mathematics.

[0192] The comprehensive evaluation unit generates a comprehensive evaluation result, for example, by referring to user-recognizable data formats pre-stored in storage unit 104, etc. The comprehensive evaluation unit may also refer to a post-processing database to generate a comprehensive evaluation result suitable for multiple evaluation results and additional information. The post-processing database is, for example, stored in storage unit 104.

[0193] The post-processing database, similar to the database described above, may also store post-processing classification information used to generate comprehensive evaluation results corresponding to multiple evaluation results and additional information. In addition to storing one or more post-processing classification information items, the post-processing database may also store, for example, multiple post-processing learning data used to generate the post-processing classification information.

[0194] Post-processing classification information is, for example, a function representing the correlation between multiple previously obtained evaluation results and additional past information (post-processing input data) and post-processing reference data associated with the post-processing input data. The post-processing reference data represents the result of a comprehensive evaluation of the user's characteristics. Post-processing classification information is generated using multiple post-processing learning data sets by treating the post-processing input data and post-processing reference data as a pair of post-processing learning data sets.

[0195] Post-processing classification information can represent, for example, a calibration model generated by analyzing the input data as explanatory variables and the reference data as the target variable using regression analysis, as described above, and based on the analysis results. Regarding post-processing classification information, in addition to periodically updating the calibration model (post-processing calibration model), it can also be generated according to additional information. Furthermore, similar to the classification information described above, post-processing classification information can also include, for example, a fully learned model generated using machine learning with multiple post-processing learning data (post-processing learned model).

[0196] In the saving step S140, the comprehensive evaluation result is saved. In the saving step S140, the comprehensive evaluation result is saved in at least one of the saving unit 104 and the server 4 by performing the same process as in the above-described embodiment.

[0197] According to this embodiment, based on the effects of the above-described embodiment, the comprehensive evaluation unit generates a comprehensive evaluation result that comprehensively evaluates the user's characteristics based on the first evaluation result, the second evaluation result, and additional information. Therefore, it is possible to achieve an evaluation that considers the user's characteristics for each evaluation result. Thus, it is possible to generate an evaluation result suitable for each user.

[0198] (Second Embodiment: A Modification of the Biological Information Processing System 100)

[0199] Next, a variation of the biological information processing system 100 in the second embodiment will be described. The difference between the second embodiment and the variation is that, in the variation, the aforementioned additional information is obtained in the generation step S120. Furthermore, descriptions of content identical to that in the embodiment described above will be omitted.

[0200] In this variation, for example, as Figure 9 As shown, in generation step S120, additional information is obtained, and a first evaluation result is generated based on the first evaluation data and the additional information. The additional information is the same as described above, for example, it is input by the user via the input unit 108, etc., and obtained by the generation unit 12, etc.

[0201] The generation unit 12 may, for example, determine the calculation method for the first evaluation data based on the content of the supplementary information. In this case, functions, etc., that differ according to each type of supplementary information are included in the classification information. Alternatively, the generation unit 12 may, for example, generate a first evaluation result based on information that combines the first evaluation data and the supplementary information.

[0202] According to this modified example, the generation unit 12 acquires additional information and generates a first evaluation result based on the first evaluation data and the additional information. Therefore, it is possible to generate a first evaluation result that considers multiple aspects of the user's characteristics based on the first evaluation data. This allows for the generation of evaluations of the user's biometric information with higher accuracy.

[0203] (Third embodiment: Biological information processing system 100)

[0204] Next, an example of the biological information processing system 100 in the third embodiment will be described. The difference between the above-described embodiment and the third embodiment is that, in the third embodiment, when generating the second evaluation result, the first evaluation result is used based on the second evaluation data. Furthermore, descriptions of contents identical to those in the above-described embodiment will be omitted.

[0205] In the biological information processing system 100 of this embodiment, for example, Figure 10As shown, in generation step S120, a second evaluation result is generated based on the first evaluation result and the second evaluation data, referring to a database. For example, generation unit 12 generates the second evaluation result after generating the first evaluation result.

[0206] The generation unit 12 may, for example, determine the calculation method for the second evaluation data based on the content of the first evaluation result. In this case, functions that differ according to each feature of the first evaluation result are included in the classification information. Alternatively, the generation unit 12 may, for example, generate the second evaluation result based on information obtained by combining the first evaluation result and the second evaluation data.

[0207] According to this embodiment, based on the effects of the above-described embodiment, the generation unit 12 generates a second evaluation result based on the first evaluation result and the second evaluation data. Therefore, a second evaluation result based on the first evaluation result can be generated. As a result, evaluations of the user's biological information can be generated with higher accuracy.

[0208] (Fourth embodiment: Biological information processing system 100)

[0209] Next, an example of the biological information processing system 100 in the fourth embodiment will be described. The difference between the above-described embodiment and the fourth embodiment is that, in the fourth embodiment, attribute-based classification information suitable for the evaluation data is selected from multiple attribute-based classification information included in the classification information. Furthermore, descriptions of content identical to that in the above-described embodiment will be omitted.

[0210] In the biological information processing device 1 of this embodiment, for example, Figure 11 As shown, generation step S120 includes a selection step S121 and a generation step S122 based on attributes. Additionally, in Figure 11 The second evaluation data and the second evaluation result are omitted from the text.

[0211] In the selection step S121, referring to the preliminary evaluation data, a specific attribute classification information (e.g., the first classification information) is selected from multiple attribute classification information. The selection step S121 can be performed, for example, by a selection unit included in the generation unit 12. The preliminary evaluation data represents features different from the first evaluation data, for example, features the same as the second evaluation data. Alternatively, the second evaluation data can also be used as the preliminary evaluation data.

[0212] In the attribute generation step S122, referring to the selected first classification information, the first organism information (e.g., the value of partial pressure of carbon dioxide in blood) corresponding to the first evaluation data is calculated, and the first evaluation result is generated. The attribute generation step S122 can be executed, for example, by the attribute generation unit included in the generation unit 12.

[0213] Multiple attribute-based classification information are calculated using different learning data. For example, when using data equivalent to the subject's acceleration pulse wave as input data for the learning data, such as according to... Figure 12 These 7 categories (A to G) are used to prepare input data and generate 7 types of attribute-based classification information.

[0214] When multiple attribute-based classification information are stored in a database, for example, the acquisition unit 11 acquires evaluation data and preliminary evaluation data corresponding to the user's acceleration pulse wave. Then, the generation unit 12 selects first classification information based on the preliminary evaluation data. Then, the generation unit 12 generates a first evaluation result corresponding to the first evaluation data based on the first classification information. Therefore, it is possible to select the classification information most suitable for the user from each attribute-based classification information.

[0215] Additionally, for example, when using data equivalent to the subject's velocity-pulse wave as input data for learning, it is also possible to follow... Figure 13 These two categories (Group 1 and Group 2) are prepared for input data to generate two types of attribute-based classification information.

[0216] Here, about Figure 12 The data shown, equivalent to acceleration pulse waves, facilitates detailed feature-based classification. However, when calculating biological information, accuracy may be reduced due to false detections of peak values. Furthermore, regarding... Figure 13 The data shown, which is equivalent to velocity pulse waves, is difficult to classify in detail based on features compared to data equivalent to acceleration pulse waves. However, due to fewer false detections of peaks, it is possible to calculate biological information with high accuracy.

[0217] Based on the above, multiple attribute-categorized information, such as those containing... Figure 12 Such data, equivalent to acceleration pulse waves, can be used as selection data for choosing specific classification information. In the learning data when generating attribute classification information, data equivalent to velocity pulse waves can also be used.

[0218] In this case, as part of the acquisition step S110, for example, the acquisition unit 11 acquires data equivalent to the velocity pulse wave based on sensor data based on the user's pulse wave as first evaluation data. Furthermore, the acquisition unit 11 acquires data equivalent to the acceleration pulse wave based on sensor data as preliminary evaluation data.

[0219] Next, as a selection step S121, for example, the generation unit 12 refers to the preliminary evaluation data, determines the selection data (first selection data) most similar to the preliminary evaluation data among a plurality of selection data including data equivalent to the acceleration pulse wave, and selects the first classification information associated with the first selection data. Then, as an attribute-based generation step S122, the generation unit 12 refers to the first classification information and generates a first evaluation result corresponding to the first evaluation data. Thus, it is possible to further improve the evaluation accuracy.

[0220] Here, an example of the data used in the above selection data and the like will be described.

[0221] For example, as Figure 12 shown, there are inflection points a to e in the acceleration pulse wave. For example, in the case of the following normalization, a method of classifying the acceleration pulse wave into 7 patterns can be used based on the magnitude relationship between the values of each inflection point and their differences. In this normalization, the maximum peak value in the acceleration pulse wave is set as point a, and each inflection point is sequentially set as point b, point c, point d, and point e starting from point a. Point a is set to 1, and the minimum value point b or point d is set to 0. First, when the value of the inflection point is b < d, it is classified into pattern A or B. If b < d and c ≥ 0.5, it is classified into A, otherwise it is classified into B. Next, when the value of the inflection point is b ≒ d, it is classified into pattern C or D. When b ≒ d and c ≒ 0, it is classified into pattern D, otherwise it is classified into pattern C. Finally, when b > d, it can be classified into any 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, and if b > c, it is classified into pattern G.

[0222] For example, the generation unit 12 determines which pattern the preliminary evaluation data, for example, conforms to Figure 12 and determines the first selection data. 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, then pattern A is used as the first selection data. Thus, it is possible to calculate the biological information with high accuracy by referring to the classification information suitable for the characteristics of the first evaluation data.

[0223] According to the present embodiment, based on the effects of the above-described embodiment, the generation unit 12 includes a selection unit that refers to the preliminary evaluation data to select the first classification information, and an attribute-based generation unit that refers to the first classification information to calculate the first evaluation result corresponding to the first evaluation data. Therefore, it is possible to calculate the first evaluation result corresponding to the first evaluation data on the basis of selecting the first classification information most suitable for the characteristics of the pulse wave. Thus, it is possible to further improve the evaluation accuracy.

[0224] Furthermore, according to this embodiment, the acquisition unit 11 acquires data equivalent to a velocity pulse wave based on a pulse wave as first evaluation data. Additionally, the acquisition unit 11 acquires data equivalent to an acceleration pulse wave based on a pulse wave as preliminary evaluation data. Therefore, acceleration pulse waves, which are easier to classify pulse wave characteristics compared to velocity pulse waves, can be used to select attribute-based classification information. Furthermore, velocity pulse waves, which are easier to calculate biological information compared to acceleration pulse waves, can be used to generate the first evaluation result. Thus, further improvement in evaluation accuracy can be achieved.

[0225] (Fourth Embodiment: First Modification of the Biological Information Processing System 100)

[0226] Next, a first variation of the biological information processing system 100 in the fourth embodiment will be described. The difference between the example of the fourth embodiment and the first variation is that, in the first variation, evaluation results are used to select classification information. Furthermore, details identical to those in the embodiments described above will be omitted.

[0227] In this variation, for example, as Figure 14 As shown, in generation step S120, based on the second evaluation result, the first classification information is selected from the classification information, and a first evaluation result corresponding to the first evaluation data is generated with reference to the first classification information. Similar to the selection step S121 described above, for example, the generation unit 12 selects a specific attribute-based classification information from multiple attribute-based classification information with reference to the second evaluation result.

[0228] The generation unit 12 selects the first classification information, for example, based on the values ​​of the organism information contained in the second evaluation result. At this time, values ​​for selection are preset among multiple attribute information.

[0229] Similar to the attribute-based generation step S122 described above, the generation unit 12 calculates the first organism information corresponding to the first evaluation data with reference to the selected first classification information, and generates the first evaluation result.

[0230] According to this modified example, the generation unit 12 selects the first classification information based on the second evaluation result, and generates a first evaluation result corresponding to the first evaluation data by referring to the first classification information. Therefore, it is possible to generate a first evaluation result corresponding to the first evaluation data based on the selection of the optimal first classification information according to the second evaluation result. As a result, the evaluation accuracy can be further improved.

[0231] (Fourth embodiment: Second variation of the biological information processing system 100)

[0232] Next, a second variation of the biological information processing system 100 in the fourth embodiment will be described. The difference between the example of the fourth embodiment and the second variation is that, in the second variation, the first classification information is selected based on the characteristics of the pulse wave. Furthermore, descriptions of contents identical to those in the embodiments described above will be omitted.

[0233] In this variation, for example, as Figure 15 As shown, in generation step S120, a first classification information is selected from the classification information based on the characteristics of the pulse wave (e.g., sensor data), and a first evaluation result corresponding to the first evaluation data is generated with reference to the first classification information. Similar to the selection step S121 described above, for example, the generation unit 12 selects a specific attribute classification information from multiple attribute classification information with reference to the sensor data.

[0234] Additionally, as a "characteristic of the pulse wave," the sensor data can also be analyzed using, for example... Figure 4 (a)~ Figure 4 The data obtained after at least a portion of the processing shown in (d). In particular, by using pulse wave data after filtering the sensor data as a feature of the pulse wave, it is possible to improve the accuracy of selecting specific attribute classification information.

[0235] The generation unit 12, for example, compares the selection data and pulse wave characteristics described above to select the first classification information. The generation unit 12 may also select the first classification information based on the half-width and relative intensity of peak values ​​contained in sensor data, etc. In this case, values ​​for selection are preset among multiple attribute information.

[0236] Similar to the attribute-based generation step S122 described above, the generation unit 12 calculates the first organism information corresponding to the first evaluation data by referring to the selected first classification information, and generates the first evaluation result.

[0237] According to this modified example, the generation unit 12 selects first classification information based on the characteristics of the pulse wave, and generates a first evaluation result corresponding to the first evaluation data by referring to the first classification information. Therefore, it is possible to generate a first evaluation result corresponding to the first evaluation data based on the selection of the optimal first classification information according to the characteristics of the pulse wave. As a result, the evaluation accuracy can be further improved.

[0238] (Fourth embodiment: Third variation of the biological information processing system 100)

[0239] Next, a third variation of the biological information processing system 100 in the fourth embodiment will be described. The difference between the example of the fourth embodiment and the third variation is that, in the third variation, the first classification information is selected based on additional information. Furthermore, descriptions of contents identical to those in the embodiments described above will be omitted.

[0240] In this variation, for example, as Figure 16 As shown, in generation step S120, a first classification information is selected from the classification information based on supplementary information, and a first evaluation result corresponding to the first evaluation data is generated with reference to the first classification information. Similar to the selection step S121 described above, for example, generation unit 12 selects a specific attribute-based classification information from multiple attribute-based classification information with reference to supplementary information. Furthermore, the supplementary information is the same as the supplementary information described above.

[0241] The generation unit 12 may, for example, select the first category information based on attribute information such as age and gender included in the supplementary information. In this case, attribute information for selection is preset among multiple attribute information.

[0242] Similar to the attribute-based generation step S122 described above, the generation unit 12 calculates the first organism information corresponding to the first evaluation data by referring to the selected first classification information, and generates the first evaluation result.

[0243] According to this modified example, the generation unit 12 selects first classification information based on additional information and generates a first evaluation result corresponding to the first evaluation data by referring to the first classification information. Therefore, it is possible to generate a first evaluation result corresponding to the first evaluation data based on the selection of the optimal first classification information according to the features of the additional information. As a result, the evaluation accuracy can be further improved.

[0244] (Fifth Embodiment: Biological Information Processing System 100)

[0245] Next, an example of the biological information processing system 100 in the fifth embodiment will be described. The difference between the above-described embodiment and the fifth embodiment is that the fifth embodiment includes a calculation step S160. Furthermore, details identical to those in the above-described embodiment will be omitted.

[0246] The organism information processing system 100 in this embodiment includes, for example, a calculation step S160. In the organism information processing system 100, the calculation step S160 is performed, for example, after the storage step S140 described above.

[0247] In calculation step S160, for example... Figure 17As shown, a comprehensive evaluation result that comprehensively evaluates the user's characteristics is generated based on multiple evaluation results (e.g., the first evaluation result and the second evaluation result) stored in the server 4 or the storage unit 104. Regarding the calculation step S160, for example, in addition to being performed by the comprehensive evaluation unit included in the generation unit 12 of the biological information processing device 1, it can also be performed by the comprehensive evaluation unit included in the server 4.

[0248] The overall evaluation results are the same as those described above. Hereinafter, as an example, we will use estimated insurance information including insurance premiums as the overall evaluation results.

[0249] The comprehensive evaluation department may, for example, generate estimated insurance information by referring to user-recognizable data pre-stored in the storage unit 104, etc. The comprehensive evaluation department may also, for example, refer to a database to generate estimated insurance information that is compatible with multiple evaluation results.

[0250] The database, similar to the one described above, can also store insurance classification information used to generate estimated insurance information corresponding to multiple evaluation results. In addition to storing one or more insurance classification pieces of information, the database can also store, for example, multiple insurance learning data used to generate the insurance classification information.

[0251] Insurance classification information, for example, is a function representing the correlation between multiple pre-obtained past evaluation results (insurance input data) and insurance reference data associated with the insurance input data. The insurance reference data contains insurance premiums corresponding to past insurance input data with proven track records. Insurance classification information is generated using multiple insurance training data sets by treating the insurance input data and insurance reference data as a pair of insurance training data.

[0252] The insurance classification information can be represented, for example, by a calibration model. This calibration model is generated through analysis using the aforementioned regression analysis, with insurance input data as explanatory variables and insurance reference data as the target variable. The calibration model (insurance calibration model) can be updated periodically, for example, using the insurance classification information. Furthermore, similar to the classification information described above, the insurance classification information can also include, for example, a fully learned model generated using machine learning with multiple insurance learning data (insurance fully learned model).

[0253] According to this embodiment, based on the effects of the above-described embodiment, the comprehensive evaluation unit generates a comprehensive evaluation result based on the first evaluation result and the second evaluation result. Therefore, it is possible to achieve an evaluation that takes into account the user's characteristics for each evaluation result. Thus, it is possible to generate an evaluation result suitable for each user. In particular, when using estimated insurance information as the comprehensive evaluation result, it is possible to suppress biases such as subjective estimations of insurance premiums by users, etc.

[0254] (Sixth Embodiment: Biological Information Processing System 100)

[0255] Next, an example of the biological information processing system 100 in the sixth embodiment will be described. The difference between the above-described embodiment and the sixth embodiment is that, in the sixth embodiment, a determination result is used. Furthermore, descriptions of contents identical to those in the above-described embodiments will be omitted.

[0256] In the biological information processing system 100 of this embodiment, for example, Figure 18 As shown, in the saving step S140, the judgment result of the user's judgment on the content of multiple evaluation results (e.g., the first evaluation result and the second evaluation result) is obtained, and the judgment result and the multiple evaluation results are associated and saved respectively. The saving step S140 can be performed by, for example, the storage unit 14 or the server 4.

[0257] For example, a user can obtain a judgment result by inputting multiple output evaluation results and biological information measured using a known measuring device through input unit 108.

[0258] The biological information processing system 100 in this embodiment may also include an update step S170. In this case, the update step S170 may be executed by the learning unit 15, or by the server 4.

[0259] The learning unit 15 updates the classification information based on the judgment results and multiple evaluation results saved in the saving step S140. The learning unit 15 updates the classification information, for example, using known techniques.

[0260] According to this embodiment, based on the effects of the above-described embodiment, the storage unit 14 or the server 4 associates and stores the determination result, the first evaluation result, and the second evaluation result respectively. Therefore, it is easy to compare each evaluation result with the determination result.

[0261] Furthermore, according to this embodiment, the learning unit 15 updates the classification information based on the judgment result, the first evaluation result, and the second evaluation result. Therefore, when the accuracy of each evaluation result decreases, it can be easily improved. As a result, the accuracy of evaluating biological information can be maintained.

[0262] (Seventh Embodiment: Biological Information Processing System 100)

[0263] Next, an example of the biological information processing system 100 in the seventh embodiment will be described. The difference between the above-described embodiment and the seventh embodiment is that, in the seventh embodiment, an electronic device 2 is used. Furthermore, descriptions of contents identical to those in the above-described embodiments will be omitted.

[0264] In the biological information processing system 100 of this embodiment, in addition to using electronic device 2 instead of biological information processing device 1 described above, biological information processing device 1 and electronic device 2 may be used separately depending on the situation.

[0265] Similar to the biological information processing device 1, the electronic device 2 refers to the aforementioned personal computer or other electronic device. The structure of the electronic device 2 may, for example, have the same characteristics as... Figure 5 The same structure as (a). Similar to the biological information processing device 1, the CPU 101 uses the RAM 103 as the working area to execute the program stored in the storage unit 104, thereby realizing the various functions of the biological information processing program for the electronic device 2.

[0266] Figure 19 The sequence of the biological information processing program used to implement the electronic device 2 is shown. The electronic device 2 has a pulse wave signal acquisition unit 60, a first extraction unit 61 and a second extraction unit 63 connected to the pulse wave signal acquisition unit 60, a first data acquisition unit 62 connected to the first extraction unit 61, a second data acquisition unit 64 connected to the second extraction unit 63, and an optimal blood glucose value calculation unit 65 connected to the first data acquisition unit 62 and the second data acquisition unit 64.

[0267] The pulse wave signal acquisition unit 60 acquires pulse wave signals transmitted from the sensor 5, server 4, and other electronic devices via the communication network 3. The pulse wave signal acquisition unit 60 outputs the acquired pulse wave signals to the first extraction unit 61 and the second extraction unit 63. Furthermore, the pulse wave signals represent the same characteristics as the sensor data described above.

[0268] The first extraction unit 61 extracts first evaluation data based on the pulse wave signal input from the pulse wave signal acquisition unit 60, referring to the first extraction conditions. The first extraction unit 61 outputs the extracted first evaluation data to the first data acquisition unit 62.

[0269] The first data acquisition unit 62 processes the first evaluation data input from the first extraction unit 61 with reference to the first processing conditions associated with the first extraction conditions, and acquires the first data. The first data acquisition unit 62 outputs the acquired first data to the optimal blood glucose value calculation unit 65.

[0270] The second extraction unit 63 extracts second evaluation data based on the pulse wave signal input from the pulse wave signal acquisition unit 60, referring to the second extraction conditions. The second extraction unit 63 outputs the extracted second evaluation data to the second data acquisition unit 64.

[0271] The second data acquisition unit 64 processes the second evaluation data input from the second extraction unit 63 with reference to the second processing conditions associated with the second extraction conditions, and acquires the second data. The second data acquisition unit 64 outputs the acquired second data to the optimal blood glucose value calculation unit 65.

[0272] The optimal blood glucose value calculation unit 65 calculates biological information that becomes the optimal value based on the first data input from the first data acquisition unit 62 and the second data input from the second data acquisition unit 64.

[0273] Furthermore, the optimal blood glucose value calculation unit 65 does not necessarily have to be installed in the electronic device 2; the first and second data can also be output as biological information.

[0274] Figure 20 The following is a specific structural example of the first extraction unit 61 and the first data acquisition unit 62. The first extraction unit 61 includes a filtering processing unit 610, a differentiating unit 611 connected to the filtering processing unit 610, a peak position calculation unit 614 connected to the filtering processing unit 610, a segmentation unit 612 connected to the differentiating unit 611, a normalization unit 613 connected to the segmentation unit 612, a peak interval averaging calculation unit 615 connected to the peak position calculation unit 614, a peak interval depiction unit 616 connected to the peak position calculation unit 614, a Fourier transform unit 617 connected to the peak position calculation unit 614, and a maximum frequency detection unit 618 connected to the Fourier transform unit 617.

[0275] The filtering processing unit 610 performs filtering processing on the acquired pulse wave signal. The filtering processing unit 610 uses, for example, a bandpass filter of 0.5 to 5 Hz, but is not limited to this. Furthermore, the filtering processing unit 610 determines an extraction method for extracting the first evaluation data from the acquired pulse wave signal based on the first extraction conditions. Based on the determined extraction method, the filtering processing unit 610 outputs the filtered pulse wave signal to at least one of the differential unit 611, the peak position calculation unit 614, and the Fourier transform unit 617.

[0276] In order to obtain a first evaluation data, the first extraction unit 61 does not necessarily have to use all the extraction methods it has, but instead uses at least one extraction method determined by the filtering processing unit 610 to extract the first evaluation data from the pulse wave signal.

[0277] Differentiation unit 611 differentiates the pulse wave signal input from filtering unit 610. When filtering unit 610 determines that differentiation is necessary, differentiation unit 611 differentiates the input pulse wave signal. Differentiation unit 611 outputs the processed pulse wave signal to segmentation unit 612.

[0278] The segmentation unit 612 divides the multiple waveform signals input from the differentiation unit 611 into segmented waveform data with integer periods. In this embodiment, the segmentation unit 612 sets the integer period to one period, but it can also set it to multiple periods. The segmentation unit 612 outputs the segmented waveform data to the normalization unit 613.

[0279] The normalization unit 613 performs normalization to unify the time width of the multiple segmented waveform signals input from the segmentation unit 612, obtains an average waveform signal that is the average of the multiple segmented waveform signals, and performs normalization by setting the maximum value of the amplitude of the average waveform signal to 1 and the minimum value to 0. The normalization unit 613 outputs the normalized average waveform signal to the regression analysis unit 620.

[0280] The normalization unit 613 can also trim the multiple segmented waveform signals input from the segmentation unit 612 with a certain time width or a certain number of samples to unify the time width of the segmented waveform signals. The processing method of the normalization unit 613 for unifying the time width is determined by the filtering processing unit 610.

[0281] When the normalization unit 613 obtains the average waveform signal, it needs multiple segmented waveform signals. The number of segmented waveform signals required is determined by the filtering processing unit 610.

[0282] The peak position calculation unit 614 calculates the peak position of the pulse wave signal input from the filtering processing unit 610 and the peak interval, which is the distance between adjacent peaks. Based on the extraction method, the peak position calculation unit 614 outputs the calculated peak interval to at least one of the peak interval averaging calculation unit 615, the peak interval depiction unit 616, and the Fourier transform unit 617.

[0283] The peak interval averaging calculation unit 615 calculates the average of the peak intervals input from the peak position calculation unit 614, divides the average peak interval by the sampling rate of the measuring device, and converts it into seconds. The peak interval averaging calculation unit 615 outputs the average value converted into seconds to the pulse processing unit 621.

[0284] The peak interval depiction unit 616 depicts a curve with the peak interval input from the peak position calculation unit 614 as the horizontal axis and the adjacent peak intervals as the vertical axis, thus obtaining a Poincaré plot of the pulse wave's peak intervals. The peak interval depiction unit 616 outputs the Poincaré plot of the pulse wave's peak intervals to the pressure depiction processing unit 622.

[0285] When the peak interval is input from the peak position calculation unit 614, the Fourier transform unit 617 performs a Fourier transform on the data after converting the peak interval into time series data. Alternatively, the Fourier transform unit 617 can also perform a Fourier transform using the pulse wave signal processed by the filtering unit 610 as input. Based on an extraction method, the Fourier transform unit 617 outputs the Fourier transformed signal to at least one of the maximum frequency detection unit 618 and the pressure Fourier processing unit 623.

[0286] The maximum frequency detection unit 61 detects the frequency, i.e., the frequency that shows a maximum value between 0.15 and 0.35 Hz, from the signal input to the Fourier transform unit 617. The maximum frequency detection unit 618 outputs the detected maximum frequency to the respiratory rate processing unit 624.

[0287] The first data acquisition unit 62 includes a regression analysis unit 620 connected to the normalization unit 613, a pulse processing unit 621 connected to the peak interval averaging unit 615, a pressure depiction processing unit 622 connected to the peak interval depiction unit 616, a pressure Fourier processing unit 623 connected to the Fourier transform unit 617, and a respiratory rate processing unit 624 connected to the maximum frequency detection unit 618.

[0288] For example, as a processing method, the first data acquisition unit 62 uses a calibration model constructed based on the correlation between measured values ​​of biological information and pre-acquired pulse wave signals to acquire the first data according to the first evaluation data.

[0289] The regression analysis unit 620 obtains, for example, blood glucose level, blood pressure, blood oxygen saturation, and blood carbon dioxide concentration as primary data based on the normalized average waveform signal input from the normalization unit 613 and a calibration model.

[0290] The regression analysis unit 620 can also store a pre-built calibration model that can be used in a general manner as general data in the first data acquisition unit 62, thereby acquiring the first data.

[0291] The pulse processing unit 621 divides the average value of the peak intervals input from the peak interval averaging calculation unit 615 by the sampling rate to convert it into seconds. The pulse processing unit 621 divides 60 seconds by the calculated number of seconds to calculate the pulse count per minute (bpm) and obtains the pulse count as the first data.

[0292] The pressure depiction processing unit 622 obtains, for example, the pressure intensity as first data based on the Poincaré chart input from the peak interval depiction unit 616. The pressure depiction processing unit 622 calculates the variance value of the Poincaré chart and estimates the pressure intensity based on the magnitude of the variance value.

[0293] The pressure Fourier processing unit 623 obtains, for example, the pressure level as first data based on the integral ratio of the signal input from the Fourier transform unit 617. Specifically, the pressure Fourier processing unit 623 calculates the PSD (power spectral density) using an autoregressive model based on the time series data of the peak intervals after Fourier transform. The region of the power spectrum from 0.05 Hz to 0.15 Hz is designated as low frequency (LF), and the region from 0.15 Hz to 0.40 Hz is designated as high frequency (HF). The pressure level is determined based on the ratio of the integral values ​​obtained by summing the intensities of the low and high frequencies.

[0294] The respiratory count processing unit 624 multiplies the frequency, which shows the maximum value, input from the maximum frequency detection unit 618 by 60 seconds to convert it into respiratory count per minute (bpm), and obtains the respiratory count as the first data.

[0295] Figure 21 A specific structural example of the second extraction unit 63 and the second data acquisition unit 64 is shown. The second extraction unit 63 includes a filtering processing unit 630, a differentiating unit 631 connected to the filtering processing unit 630, a peak position calculation unit 634 connected to the filtering processing unit 630, a segmentation unit 632 connected to the differentiating unit 631, a normalization unit 633 connected to the segmentation unit 632, a peak interval averaging calculation unit 635 connected to the peak position calculation unit 634, a peak interval depiction unit 636 connected to the peak position calculation unit 634, a Fourier transform unit 637 connected to the peak position calculation unit 634, and a maximum frequency detection unit 638 connected to the Fourier transform unit 637. The second extraction unit 63 is composed of the same parts as the first extraction unit 61, but while the first extraction unit 61 refers to the first extraction condition, the second extraction unit 63 refers to the second extraction condition instead of the first extraction condition.

[0296] The filtering processing unit 630 performs filtering processing on the acquired pulse wave signal. The filtering processing unit 630 uses, for example, a bandpass filter of 0.5 to 5 Hz, but is not limited to this. Furthermore, the filtering processing unit 630 determines a method for extracting the first evaluation data from the acquired pulse wave signal based on the second extraction conditions. Based on the determined extraction method, the filtering processing unit 630 outputs the filtered pulse wave signal to at least one of the differential unit 631, the peak position calculation unit 634, and the Fourier transform unit 637.

[0297] In order to obtain a second evaluation data, the second extraction unit 63 does not necessarily have to use all the extraction methods it has, but instead uses at least one extraction method determined by the filtering processing unit 630 to extract the second evaluation data from the pulse wave signal.

[0298] Differentiation unit 631 differentiates the pulse wave signal input from filtering unit 630. When filtering unit 630 determines that differentiation is required, differentiation unit 631 differentiates the input pulse wave signal. Differentiation unit 631 outputs the differentiated pulse wave signal to segmentation unit 632.

[0299] The segmentation unit 632 divides the multiple waveform signals input from the differentiation unit 631 into segmented waveform data with integer periods. In this embodiment, the segmentation unit 632 sets the integer period to one period, but it can also set it to multiple periods. The segmentation unit 632 outputs the segmented waveform data to the normalization unit 633.

[0300] The normalization unit 633 performs normalization to unify the time width of the multiple segmented waveform signals input from the segmentation unit 632, obtains an average waveform signal that is the average of the multiple segmented waveform signals, and performs normalization by setting the maximum value of the amplitude of the average waveform signal to 1 and the minimum value to 0. The normalization unit 633 outputs the normalized average waveform signal to the regression analysis unit 640.

[0301] The normalization unit 633 can also trim the multiple segmented waveform signals input from the segmentation unit 632 with a certain time width or a certain number of samples to unify the time width of the segmented waveform signals. The processing method of the normalization unit 633 for unifying the time width is determined by the filtering processing unit 630.

[0302] When the normalization unit 633 obtains the average waveform signal, it needs multiple segmented waveform signals. The number of segmented waveform signals required is determined by the filtering processing unit 630.

[0303] The peak position calculation unit 634 calculates the peak position of the pulse wave signal input from the filtering processing unit 630 and the peak interval, which is the distance between adjacent peaks. Based on the extraction method, the peak position calculation unit 634 outputs the calculated peak interval to at least one of the peak interval averaging calculation unit 635, the peak interval depiction unit 636, and the Fourier transform unit 637.

[0304] The peak interval averaging calculation unit 635 calculates the average of the peak intervals input from the peak position calculation unit 634, and divides the average peak interval by the sampling rate of the measuring device to convert it into seconds. The peak interval averaging calculation unit 635 outputs the average value converted into seconds to the pulse processing unit 641.

[0305] The peak interval depiction unit 636 depicts a curve with the peak interval input from the peak position calculation unit 634 as the horizontal axis and the adjacent peak intervals as the vertical axis, thus obtaining a Poincaré plot of the pulse wave's peak intervals. The peak interval depiction unit 636 outputs the Poincaré plot of the pulse wave's peak intervals to the pressure depiction processing unit 642.

[0306] When the peak interval is input from the peak position calculation unit 634, the Fourier transform unit 637 performs a Fourier transform on the data after converting the peak interval into time series data. Alternatively, the Fourier transform unit 637 can also perform a Fourier transform using the pulse wave signal processed by the filtering unit 630 as input. Based on an extraction method, the Fourier transform unit 637 outputs the Fourier transformed signal to at least one of the maximum frequency detection unit 638 and the pressure Fourier processing unit 643.

[0307] The maximum frequency detection unit 63 detects the frequency, i.e., the frequency that shows a maximum value between 0.15 and 0.35 Hz, from the signal input to the Fourier transform unit 637. The maximum frequency detection unit 638 outputs the detected maximum frequency to the respiratory rate processing unit 644.

[0308] The second data acquisition unit 64 includes a regression analysis unit 640 connected to the normalization unit 633, a pulse processing unit 641 connected to the peak interval average calculation unit 635, a pressure depiction processing unit 642 that obtains pressure intensity based on the curve drawn by the peak interval depiction unit 636, a pressure Fourier processing unit 643 connected to the Fourier transform unit 637, and a respiratory rate processing unit 644 connected to the maximum frequency detection unit 638.

[0309] The second data acquisition unit 64 processes the second evaluation data extracted by the second extraction unit 63 with reference to the second processing conditions associated with the second extraction conditions, and obtains the second data.

[0310] For example, as a processing method, the second data acquisition unit 64 uses a calibration model constructed based on the correlation between measured values ​​based on biological information and pre-acquired pulse wave signals to acquire second data based on the second evaluation data.

[0311] The regression analysis unit 640 obtains, for example, blood glucose level, blood pressure, blood oxygen saturation, and blood carbon dioxide concentration as second data based on the normalized average waveform signal input from the normalization unit 633 and a calibration model.

[0312] The regression analysis unit 640 can also store a pre-built calibration model that can be used in a general manner as general data in the second data acquisition unit 64, thereby acquiring the second data.

[0313] The pulse processing unit 641 divides the average value of the peak intervals input from the peak interval averaging calculation unit 635 by the sampling rate to convert it into seconds. The pulse processing unit 641 divides 60 seconds by the calculated number of seconds to calculate the pulse count per minute (bpm) and obtains the pulse count as the second data.

[0314] The pressure depiction processing unit 642 obtains, for example, the pressure intensity as second data based on the Poincaré chart input from the peak interval depiction unit 636. The pressure depiction processing unit 642 calculates the variance value of the Poincaré chart and estimates the pressure intensity based on the magnitude of the variance value.

[0315] The pressure Fourier processing unit 643 obtains the pressure level as second data based on the integral ratio of the signal input from the Fourier transform unit 637, for example.

[0316] The respiratory count processing unit 644 multiplies the frequency, which shows the maximum value, input from the maximum frequency detection unit 638 by 60 seconds to convert it into respiratory count per minute (bpm), and obtains the respiratory count as the second data.

[0317] The first and second data obtained by the biological information processing system 100 include, for example, at least one of the following: blood pressure, blood glucose level, blood oxygen saturation, blood carbon dioxide concentration, pulse rate, respiratory rate, pressure level, vascular age, degree of diabetes, etc.

[0318] Next, an example of the operation of the biological information processing system 100 in this embodiment will be described. Figure 22 This is a flowchart illustrating an example of the operation of the biological information processing system 100 in this embodiment.

[0319] In the pulse wave signal acquisition step S10, the acquisition unit 50 acquires the pulse wave signal and outputs it to the communication I / F 51 via the internal bus 54. Alternatively, the bio-information processing system 100 can output the pulse wave signal recorded in the memory 52 to the communication I / F 51 via the internal bus 54, instead of the pulse wave signal acquired by the acquisition unit 50. For example, an FBG sensor can be used to acquire the pulse wave signal.

[0320] Next, the communication I / F 51, which has received the pulse wave signal from the acquisition unit 50, sends the pulse wave signal to the pulse wave signal acquisition unit 60 via the communication network 3. Alternatively, at this time, the pulse wave signal stored in the server 4 can be sent to the pulse wave signal acquisition unit 60 instead of the pulse wave signal acquired by the acquisition unit 50.

[0321] Next, the pulse wave signal acquisition unit 60, which has transmitted the pulse wave signal via the communication network 3, outputs the pulse wave signal to the first extraction unit 61 and the second extraction unit 63.

[0322] Next, in the first extraction condition determination step S11, the first extraction unit 61 inputs the pulse wave signal input from the pulse wave signal acquisition unit 60 to the filtering processing unit 610. Then, the filtering processing unit 610 determines the extraction method to be performed on the input pulse wave signal with reference to the first extraction condition.

[0323] The filtering processing unit 610 determines the extraction method performed by the first extraction unit 61 on the pulse wave signal from the first extraction conditions based on the state of the acquired pulse wave signal, the acquired biological information, and external factors. External factors include, for example, at least one of the following: user information such as age, gender, medical history, lifestyle habits, medication information, degree of arteriosclerosis, health status, and genetic information; and environmental information such as temperature and humidity. For example, if the desired biological information is blood glucose levels, the filtering processing unit 610 sends a command to the first extraction unit 61, causing the pulse wave signal to be processed by the filtering processing unit 610 and then output to the differential unit 611. The first extraction conditions are a data set containing a list of extraction methods for extracting first evaluation data from the pulse wave signal. The data set of the first extraction conditions may contain multiple extraction methods or a pre-determined extraction method.

[0324] Next, in the first extraction step S12, the first extraction unit 61 extracts first evaluation data from the pulse wave signal input from the pulse wave signal acquisition unit 60. The first evaluation data is waveform data extracted from the pulse wave signal by the first extraction unit 61 for obtaining biological information by the first data acquisition unit 62. The first evaluation data is, for example, waveform data obtained by processing the pulse wave signal into a waveform of one cycle through at least one of the following processes: filtering, differentiation, normalization, and averaging. In this embodiment, as an example, a biological information processing system 100 for obtaining blood glucose values ​​will be described.

[0325] First, after filtering the pulse wave signal input from the pulse wave signal acquisition unit 60, the filtering processing unit 610 obtains, for example, blood glucose value as biological information, and outputs the pulse wave signal to the differential unit 611 based on the extraction method determined in the first extraction condition determination step S11.

[0326] Next, the differentiation unit 611 determines whether to differentiate the pulse wave signal input from the filtering processing unit 610 based on the extraction method determined in the first extraction condition determination step S11, and outputs the pulse wave signal to the segmentation unit 612 after processing.

[0327] The reason for determining whether the differentiation unit 611 should differentiate the pulse wave signal based on the extraction method determined in the first extraction condition determination step S11 is that the characteristics of the first evaluation data obtained by differentiating the pulse wave signal versus not differentiating it will differ, and the first evaluation data suitable for the desired first data is obtained using the first data. Furthermore, "differentiating the pulse wave signal" means extracting the pulse wave signal as an acceleration pulse wave, and "not differentiating the pulse wave signal" means extracting the pulse wave signal as a velocity pulse wave.

[0328] Next, the segmentation unit 612 divides the multiple waveform signals input from the differentiation unit 611 into segmented waveform data of one cycle for averaging. Then, the segmentation unit 612 outputs the segmented waveform data to the normalization unit 613.

[0329] The normalization unit 613 normalizes the horizontal axis to unify the time width of the multiple segmented waveform signals input from the segmentation unit 612, obtains an average waveform signal that is the average of the multiple segmented waveform signals, and performs vertical axis normalization by setting the maximum value of the average waveform signal to 1 and the minimum value to 0. At this time, the first extraction unit 61 acquires the average waveform signal as the first evaluation data. Then, the normalization unit 613 outputs the normalized average waveform signal to the first data acquisition unit 62.

[0330] The reason for normalizing the horizontal axis by the normalization unit 613 to unify the time width is that there are significant differences at the end of the pulse wave, so this part is deleted and the main part of the pulse wave is used as the analysis object. Furthermore, the reason for normalizing the vertical axis by setting the maximum value of the average waveform signal to 1 and the minimum value to 0 is to average the deviation of the pressing pressure when the FBG sensor is installed at the measurement site, and the deviation of the measurement data caused by the positional shift of the FBG sensor during measurement, to suppress noise caused by deviations during measurement, and to improve the accuracy of the correlation between the pulse wave signal and the measured values ​​of biological information.

[0331] Next, the biological information processing system 100 moves to the first data acquisition step S13, and with reference to the first processing conditions associated with the first extraction conditions, processes the first evaluation data input from the first extraction unit 61 using the first data acquisition unit 62 to acquire the first data.

[0332] The first processing condition is a data set containing a method for processing the first evaluation data input from the first extraction unit 61 by the first data acquisition unit 62, and this method is associated with the first extraction condition. The first data acquisition unit 62 determines the processing method from the first processing condition. The data set of the first processing condition may contain multiple processing methods, or it may contain a pre-determined processing method.

[0333] For example, when the first data acquisition unit 62 processes the average waveform signal obtained by the above extraction method as the first evaluation data, the first data acquisition unit 62 decides to output the average waveform signal to the regression analysis unit 620, and decides the processing method to be performed on the average waveform signal by the regression analysis unit 620.

[0334] The regression analysis unit 620, which receives the average waveform signal from the first extraction unit 61, uses a calibration model representing the correlation between the measured value and the pulse wave signal to obtain, for example, the blood glucose value as the first data based on the average waveform signal and outputs it.

[0335] Calibration models are constructed, for example, by using regression analysis with a pre-measured average waveform signal as the explanatory variable and the measured value of biological information as the target variable, and based on the analysis results. Regarding calibration models, pre-constructed calibration models that can be used in a general manner can also be stored in storage or similar locations to measure the first data. The construction of calibration models is sometimes necessary, for example, in cases of periodic calibration or when the user changes their settings.

[0336] Furthermore, when estimating biological information such as blood carbon dioxide concentration and other values ​​that are difficult to observe and for which data on outliers are difficult to collect, the regression analysis unit 620 can also estimate outliers in biological information based on the deviation between the average waveform signal input from the first extraction unit 61 and the calibration model.

[0337] The regression analysis unit 620 has multiple calibration models. The processing method determined in the first data acquisition unit 62, based on the first processing conditions, determines which calibration model to use for the input average waveform data. For example, if the filtering processing unit 610 selects blood glucose value as the desired biological information, the first data acquisition unit 62 determines that the average waveform signal input from the first extraction unit 61 will be output to the regression analysis unit 620. The regression analysis unit 620 uses a calibration model to acquire the first data based on the input average waveform data. This calibration model is constructed by performing regression analysis with pre-measured pulse wave waveform data as the explanatory variable and the measured blood glucose value as the target variable, and based on the analysis results.

[0338] Furthermore, for example, if the filtering processing unit 610 determines the signal extraction conditions based on the user's age as an external factor, the first data acquisition unit 62 determines, by referring to the first processing conditions associated with the first extraction conditions, that the average waveform signal input from the first extraction unit 61 will be output to the regression analysis unit 620. The regression analysis unit 620 acquires the first data based on the input average waveform data using a calibration model that is constructed by performing regression analysis with pre-measured pulse wave waveform data as the explanatory variable and measured blood glucose values ​​of users close to the user's age as the target variable. Thus, a processing method suitable for the extraction method can be determined.

[0339] Next, in the second extraction condition determination step S14, the second extraction unit 63 inputs the pulse wave signal input from the pulse wave signal acquisition unit 60 to the filtering processing unit 630. Then, the filtering processing unit 630 determines the extraction method to be performed on the input pulse wave signal with reference to the second extraction condition.

[0340] The second extraction condition is a data set containing a list of extraction methods for extracting the second evaluation data from the pulse wave signal. The data set for the second extraction condition may contain multiple extraction methods, or it may contain a pre-determined extraction method. Furthermore, the second extraction condition may be the same as the first extraction condition.

[0341] The filtering processing unit 630 determines the extraction method performed by the second extraction unit 63 on the pulse wave signal from the second extraction conditions based on the state of the desired pulse wave signal, the desired biological information, and external factors. External factors include, for example, user information such as age, gender, medical history, lifestyle habits, health status, and genetic information, as well as environmental information such as temperature and humidity. Furthermore, the filtering processing unit 630 can also determine the extraction method performed by the second extraction unit 63 on the pulse wave signal based on the content of the first data obtained by the first data acquisition unit 62. For example, if the accuracy of the blood glucose value in the first data obtained by the first data acquisition unit 62 is low, the filtering processing unit 630 determines an extraction method not performed by the first extraction unit 61. Furthermore, for example, if blood glucose value is obtained as the first data, in order to obtain blood pressure as the second data, the filtering processing unit 630 can also determine the following extraction conditions: the normalization unit 633 adjusts the waveform signal with a certain number of samples to uniformly divide the time width. Therefore, second data that matches the changes and characteristics of the first data can be obtained, resulting in a highly accurate evaluation result. Furthermore, the filtering processing unit 630 can determine the processing method implemented by the second data acquisition unit 64 on the second evaluation data based on the content of the first data acquired by the first data acquisition unit 62. Specifically, this method can be the same as the method described later where the second data acquisition unit 64 determines the processing method implemented on the second evaluation data based on the content of the first data acquired by the first data acquisition unit 62.

[0342] Furthermore, the filtering processing unit 630 can also determine the extraction method implemented by the second extraction unit 63 for the pulse wave signal based on the extraction method of the first evaluation data obtained by the first data acquisition unit 62. For example, when acquiring the first evaluation data, if differentiation is performed by the differentiation unit 611, the filtering processing unit 630 decides that the differentiation unit 631 should not perform differentiation. Thus, the optimal second data can be obtained according to the change of the first data.

[0343] Next, in the second extraction step S15, the second extraction unit 63 extracts second evaluation data from the pulse wave signal input from the pulse wave signal acquisition unit 60. The second evaluation data is waveform data extracted from the pulse wave signal by the second extraction unit 63 for processing by the second data acquisition unit 64. The second evaluation data is, for example, waveform data obtained by processing the pulse wave signal into a waveform of one cycle through at least one of the following processes: filtering, differentiation, normalization, and averaging.

[0344] First, after the filtering processing unit 630 performs filtering processing on the pulse wave signal input from the pulse wave signal acquisition unit 60, for example, to obtain blood glucose value as biological information, it outputs the pulse wave signal to the differential unit 631 based on the extraction method determined in the second extraction condition determination step S14.

[0345] Next, the differentiation unit 631 determines whether to differentiate the pulse wave signal input from the filtering unit 630 based on the extraction method determined in the second extraction condition determination step S14, and outputs the pulse wave signal to the segmentation unit 632 after processing.

[0346] Next, the segmentation unit 632 divides the multiple waveform signals input from the differentiation unit 631 into segmented waveform data for one cycle. Then, the segmentation unit 632 outputs the segmented waveform data to the normalization unit 633.

[0347] The normalization unit 633 performs normalization to unify the time width of the multiple segmented waveform signals input from the segmentation unit 632, obtaining an average waveform signal that is the average of the multiple segmented waveform signals. Normalization is then performed by setting the maximum value of the average waveform signal to 1 and the minimum value to 0. At this time, the second extraction unit 63 acquires the average waveform signal as second evaluation data. Then, the normalization unit 633 outputs the normalized average waveform signal to the second data acquisition unit 64.

[0348] Next, the biological information processing system 100 moves to the second data acquisition step S16, and with reference to the second processing conditions associated with the second extraction conditions, the second data acquisition unit 64 processes the second evaluation data input from the second extraction unit 63 to acquire the second data.

[0349] The second processing condition is a data set containing a method for processing the second evaluation data input from the second extraction unit 63 by the second data acquisition unit 64, and this method is associated with the second extraction condition. The second data acquisition unit 64 determines the processing method from the second processing condition. The data set of the second processing condition may contain multiple processing methods, or it may contain a pre-determined processing method. Furthermore, the second processing condition may be the same as the first processing condition.

[0350] Furthermore, the second data acquisition unit 64 can also determine the processing method for the second evaluation data implemented by the second data acquisition unit 64 based on the processing method of the first data. For example, if a processing method is implemented that uses a calibration model representing the correlation between measured values ​​from younger users and pulse wave signals to acquire the first data based on the average waveform signal, the second data acquisition unit 64 can also use a calibration model more suitable for the user to acquire the second data based on the average waveform signal, according to the aforementioned processing method. Thus, multiple biological information obtained through different processing methods can be acquired, enabling more comprehensive and accurate evaluation.

[0351] Furthermore, the second data acquisition unit 64 can also determine the processing method corresponding to the content of the first data as the processing method for the second evaluation data implemented by the second data acquisition unit 64, referring to the second processing conditions. For example, if it is determined from the first data that the user has a tendency to hypoglycemia, the second data acquisition unit 64 can determine the following processing method as the processing method for the second evaluation data implemented by the second data acquisition unit 64: using a calibration model that represents the correlation between the measured blood glucose value of the user with hypoglycemia and the pulse wave signal to obtain the blood glucose value. It is conceivable that the obtained blood glucose value may have a large error. Therefore, by classifying the blood glucose range into hypoglycemia, normal blood glucose, hyperglycemia, and hyperglycemia based on the content of the first data, and using a calibration model that matches the blood glucose range, the accuracy of the blood glucose value obtained as the second data can be greatly improved. In addition, it is also possible to determine whether the user has diabetes based on the content of the first data, and combine this result to determine the processing method for the second evaluation data implemented by the second data acquisition unit 64. Thus, by determining the processing of the second evaluation data in accordance with the changes in the first data, a higher accuracy evaluation can be performed.

[0352] For example, the regression analysis unit 620 can also classify the processing methods according to the content of the calibration model used in the process of obtaining the first data from the average waveform signal. For example, the processing method of obtaining the first data from the average waveform signal using a calibration model that represents the correlation between measured values ​​and pulse wave signals from users in the younger age group is a different processing method than the processing method of obtaining the first data from the average waveform signal using a calibration model that represents the correlation between measured values ​​and pulse wave signals from users in other age groups.

[0353] Furthermore, the second data acquisition unit 64 can also determine the processing method for the second evaluation data implemented by the second data acquisition unit 64 by referring to the classification mode of the first evaluation data, which uses the result of classifying the first evaluation data as the first data. Thus, for example, after classifying the signal features using easily classifiable acceleration pulse waves, it is possible to perform high-precision evaluation using velocity pulse waves that can suppress false detections. Therefore, it is possible to determine the processing method for the second evaluation data that is more suitable for the user by following changes in the first data.

[0354] A classification pattern is a classification table used to classify signals into two or more groups based on the characteristics of their waveforms. For example, using... Figure 12 The classification pattern of acceleration pulse waves is shown.

[0355] When the first data acquisition unit 62 receives first evaluation data based on the acceleration pulse wave after differentiating the pulse wave signal by the differentiating unit 611, it determines that the first evaluation data, for example, meets the following criteria: Figure 12 The classification mode is determined by which mode is selected. For example, if the inflection point b of the first evaluation data is less than the inflection point d and the inflection point c ≥ 0.5, then mode A is selected as the classification mode for the first evaluation data.

[0356] The appropriate calibration model varies depending on each classification of the acceleration pulse wave. Therefore, the second data acquisition unit 64 determines the processing method for the second evaluation data by referring to the classification pattern of the first evaluation data after classifying the first evaluation data as the first data, thereby enabling a processing method suitable for each user. For example, when the first evaluation data is classified as pattern A as the first data, the second data acquisition unit 64 determines that the regression analysis unit 640 uses a calibration model to acquire the second data based on the input average waveform data. This calibration model is constructed by performing regression analysis with the pre-measured waveform data of pattern A of the pulse wave as the explanatory variable and the measured value of biological information as the target variable, and based on the analysis result.

[0357] For example, it can also be used Figure 13 The classification mode of the velocity pulse wave shown is used instead of the classification mode of the acceleration pulse wave described above. By using the classification mode as described above, the first data acquisition unit 62 can also determine the classification mode when it is inputted with first evaluation data based on the velocity pulse wave that has not been differentiated by the differential unit 611.

[0358] Next, in the optimal blood glucose value calculation step S17, the first data obtained by the first data acquisition unit 62 and the second data obtained by the second data acquisition unit 64 are input to the optimal blood glucose value calculation unit 65, and optimal biological information is obtained based on the first data and the second data. For example, the blood glucose value obtained as the first data is set as the first blood glucose value, and the blood glucose value obtained as the second data is set as the second blood glucose value, and the optimal blood glucose value is calculated based on the first blood glucose value and the second blood glucose value. Regarding the method for calculating the biological information that becomes the optimal value, for example, the first data and the second data may be weighted according to their respective measurement accuracy, and the biological information that becomes the optimal value may be calculated based on the weighting of multiple first data and multiple second data. In addition, as other examples, the following methods can be cited: setting the blood glucose value obtained by other sensors as the reference value, generating a plotting curve of the reference value and the first and second data, and outputting the data that shows a good value on the error grid in the two plotting curves; obtaining multiple first and second data respectively, and outputting the data with small deviation; evaluating whether the first and second data are within the range within a preset allowable range, and outputting the data that are within the range.

[0359] Regarding the pulse wave waveform signal, the numerical values ​​of biological information that can be obtained can sometimes be biased depending on the extraction conditions and processing methods. For example, when comparing a pulse wave processed using a differentiated extraction condition with a pulse wave processed using an undifferentiated extraction condition, the numerical values ​​of biological information obtained may be biased. That is, in order to obtain sufficiently accurate biological information from the pulse wave waveform signal, it is necessary to evaluate the pulse wave waveform signal from multiple perspectives based on multiple biological information outputs simultaneously after various extraction conditions.

[0360] On the other hand, existing technologies do not disclose the simultaneous output of multiple biological information from a single pulse wave waveform signal through various extraction conditions. Therefore, the biological information obtainable based on extraction conditions in existing technologies may be biased, potentially resulting in insufficient accuracy.

[0361] In contrast, the biological information processing system 100 in this embodiment includes, for example, a pulse wave signal acquisition step S10 for acquiring a velocity pulse wave as a pulse wave signal, a first extraction step S12 for extracting first evaluation data based on the pulse wave signal according to a first extraction condition, a first data acquisition step S13 for acquiring first data corresponding to the first evaluation data according to a first processing condition associated with the first extraction condition, a second extraction step S15 for extracting second evaluation data based on the pulse wave signal according to a second extraction condition, and a second data acquisition step S16 for acquiring second data corresponding to the second evaluation data according to a second processing condition associated with the second extraction condition. Furthermore, in the second extraction step S15, any one of the multiple extraction methods included in the second extraction condition to be referenced, or any one of the multiple processing methods included in the second processing condition, is determined based on the first data.

[0362] That is, according to this embodiment, first evaluation data based on the pulse wave signal is extracted by referring to the first extraction step S12 of the first extraction condition, and the first data corresponding to the first evaluation data is obtained by referring to the first data acquisition step S13 of the first processing condition associated with the first extraction condition. Furthermore, the biological information processing system 100 extracts second evaluation data based on the pulse wave signal by referring to the second extraction step S15 of the second extraction condition, and the second data corresponding to the second evaluation data is obtained by referring to the second data acquisition step S16 of the second processing condition associated with the second extraction condition. Thus, multiple biological information samples, for which different extraction and processing methods have been implemented, can be simultaneously obtained based on a single input pulse wave signal. Using this multiple biological information samples, a high-precision evaluation result can be obtained by evaluating multiple aspects of the biological information using a single pulse wave signal.

[0363] Furthermore, according to this embodiment, in the second extraction step S15, based on the first data obtained through the first data acquisition step S13, it is determined which method among the multiple extraction methods included in the second extraction conditions to be referenced, or which method among the multiple processing methods included in the aforementioned second processing conditions, will be used. Therefore, the most suitable extraction method for the pulse wave signal can be determined based on the content of the first data, and second data matching the changes and characteristics of the first data can be obtained. Thus, a high-precision evaluation result matching the user's characteristics can be obtained.

[0364] Furthermore, according to this embodiment, in the second extraction step S15, any one of the multiple extraction methods included in the second extraction conditions to be referenced is determined based on the extraction method of the first evaluation data obtained in the first extraction step S12. Therefore, in the second extraction step S15, a more suitable extraction method can be determined based on the extraction method of the first evaluation data, enabling the acquisition of multiple biological information with different extraction methods more suitable for pulse wave signals, and obtaining higher-precision evaluation results through multi-faceted evaluation.

[0365] Furthermore, according to this embodiment, in the second data acquisition step S16, any one of the multiple processing methods included in the second processing conditions to be referenced is determined based on the processing method of the first data determined in the first data acquisition step S13. Therefore, in the second data acquisition step S16, the optimal processing method for the second evaluation data can be determined based on the processing method of the first data, enabling the acquisition of multiple biological information with different processing methods more suitable for pulse wave signals, and obtaining more accurate evaluation results through more comprehensive evaluation.

[0366] Furthermore, according to this embodiment, in the first extraction step S12, the pulse wave signal is differentiated to extract the first evaluation data. In the first data acquisition step S13, the classification result of the first evaluation data is obtained as the first data, referring to the classification mode of the first evaluation data. In the second extraction step S15, the pulse wave signal is not differentiated, and the second evaluation data is extracted. Thus, acceleration pulse waves suitable for classification can be classified as the first evaluation data, and velocity pulse waves that can suppress false detections can be processed as the second evaluation data. A higher accuracy evaluation result can be obtained through multi-faceted evaluation.

[0367] (Eighth embodiment: Biological information processing system 100)

[0368] Next, an example of the biological information processing system 100 in the eighth embodiment will be described. The difference between the above-described embodiment and the eighth embodiment is that in the eighth embodiment, the electronic device 2 performs different processing. Furthermore, descriptions of contents identical to those in the above-described embodiments will be omitted.

[0369] Figure 23The sequence of the biological information processing program for implementing the electronic device 2 is shown. The electronic device 2 includes a pulse wave signal acquisition unit 60a, a classification data extraction unit 61a and an evaluation data extraction unit 65a connected to the pulse wave signal acquisition unit 60a, a first mode selection unit 62a connected to the classification data extraction unit 61a, a second mode selection unit 66a connected to the evaluation data extraction unit 65a, a first processing selection unit 63a connected to the first mode selection unit 62a and the second mode selection unit 66a, and a blood glucose value acquisition unit 64a connected to the first processing selection unit 63a and the evaluation data extraction unit 65a.

[0370] The pulse wave signal acquisition unit 60a acquires pulse wave signals transmitted from the sensor 5, server 4, and other electronic devices 2 via the communication network 3. The pulse wave signal acquisition unit 60a outputs the acquired pulse wave signals to the classification data extraction unit 61a and the evaluation data extraction unit 65a. Furthermore, when it is not necessary for the evaluation data extraction unit 65a to extract the pulse wave signal, the pulse wave signal acquisition unit 60a may also output the pulse wave signal to the blood glucose value acquisition unit 64a.

[0371] The classification data extraction unit 61a extracts classification data based on the pulse wave signal input from the pulse wave signal acquisition unit 60a, referring to the extraction conditions. The classification data extraction unit 61a outputs the extracted classification data to the first mode selection unit 62a.

[0372] The first mode selection unit 62a selects the first mode corresponding to the classification data input from the classification data extraction unit 61a, referring to the first classification mode. The first mode selection unit 62a outputs the selected first mode to the first processing selection unit 63a.

[0373] The evaluation data extraction unit 65a extracts evaluation data based on the pulse wave signal input from the pulse wave signal acquisition unit 60a, referring to the evaluation extraction conditions. The evaluation data extraction unit 65a outputs the extracted evaluation data to the second mode selection unit 66a and the blood glucose value acquisition unit 64a.

[0374] The second mode selection unit 66a refers to the second classification mode and selects the second mode corresponding to the evaluation data input from the evaluation data extraction unit 65a. The second mode selection unit 66a outputs the selected second mode to the first processing selection unit 63a.

[0375] The first processing selection unit 63a refers to a pre-acquired processing mode and selects a first processing mode based on the first mode input from the first mode selection unit 62a and the second mode input from the second mode selection unit 66a. The first processing selection unit 63a outputs the selected first processing mode to the blood glucose value acquisition unit 64a.

[0376] The blood glucose value acquisition unit 64a obtains the blood glucose value corresponding to the evaluation data input from the evaluation data extraction unit 65a by referring to the first processing input from the first processing selection unit 63a.

[0377] Figure 24 A specific structural example of the classification data extraction unit 61a is shown. The classification data extraction unit 61a has a filtering processing unit 610a connected to the pulse wave signal acquisition unit 60a, a differentiating unit 611a connected to the filtering processing unit 610a, a segmentation unit 612a connected to the differentiating unit 611a, and a normalization unit 613a connected to the segmentation unit 612a.

[0378] In order to obtain one categorical data, the categorical data extraction unit 61a does not necessarily have to use all the extraction methods available in the categorical data extraction unit 61a, but instead uses at least one extraction method determined by the extraction conditions to extract categorical data from the pulse wave signal.

[0379] The filtering unit 610a performs filtering processing on the acquired pulse wave signal. The filtering unit 610a uses, for example, a bandpass filter of 0.5 to 5 Hz, but is not limited to this. Furthermore, the filtering unit 610a determines the extraction method for extracting classification data from the acquired pulse wave signal based on extraction conditions. The filtering unit 610a outputs the filtered pulse wave signal to the differentiating unit 611a.

[0380] Differentiation unit 611a differentiates the pulse wave signal input from filtering unit 610a. When filtering unit 610a determines that differentiation is necessary, differentiation unit 611a performs differentiation on the input pulse wave signal. Differentiation unit 611a outputs the processed pulse wave signal to segmentation unit 612a.

[0381] The segmentation unit 612a divides the multiple waveform signals input from the differentiation unit 611a into segmented waveform data with integer periods. In this embodiment, the segmentation unit 612a sets the integer period to one period, but it can also set it to multiple periods. The segmentation unit 612a outputs the segmented waveform data to the normalization unit 613a.

[0382] The normalization unit 613a performs normalization to unify the time width of the multiple segmented waveform signals input from the segmentation unit 612a, obtains an average waveform signal that is the average of the multiple segmented waveform signals, and performs normalization by setting the maximum value of the amplitude of the average waveform signal to 1 and the minimum value to 0, thereby obtaining classification data. The normalization unit 613a outputs the obtained classification data to the first mode selection unit 62a.

[0383] The normalization unit 613a can also trim the multiple segmented waveform signals input from the segmentation unit 612a with a certain time width or a certain number of samples to unify the time width of the segmented waveform signals. The processing method of the normalization unit 613a for unifying the time width is determined by the filtering processing unit 610a.

[0384] When the normalization unit 613a obtains the average waveform signal, it needs multiple segmented waveform signals. The number of segmented waveform signals required is determined by the filtering processing unit 610a.

[0385] The first mode selection unit 62a classifies the classification data input from the normalization unit 613a into the first mode by referring to the pre-acquired first classification mode. The first mode selection unit 62a outputs the classified first mode to the first processing selection unit 63a.

[0386] Figure 25 A specific structural example of the evaluation data extraction unit 65a is shown. The evaluation data extraction unit 65a includes a filtering processing unit 650a connected to the pulse wave signal acquisition unit 60a, a differentiating unit 651a connected to the filtering processing unit 650a, a segmentation unit 652a connected to the differentiating unit 651a, and a normalization unit 653a connected to the segmentation unit 652a. The evaluation data extraction unit 65a has the same structure as the classification data extraction unit 61a, but while the classification data extraction unit 61a refers to extraction conditions, the evaluation data extraction unit 65a refers to evaluation extraction conditions instead of extraction conditions.

[0387] The filtering processing unit 650a performs filtering processing on the acquired pulse wave signal. The filtering processing unit 650a uses, for example, a bandpass filter of 0.5 to 5 Hz, but is not limited to this. Furthermore, the filtering processing unit 650a determines the extraction method for extracting evaluation data from the acquired pulse wave signal based on the extraction conditions for evaluation. The filtering processing unit 650a outputs the filtered pulse wave signal to the differentiating unit 651a.

[0388] Differentiation unit 651a differentiates the pulse wave signal input from filtering unit 650a. When filtering unit 650a determines that differentiation is necessary, differentiation unit 651a performs differentiation on the input pulse wave signal. Differentiation unit 651a outputs the processed pulse wave signal to segmentation unit 652a.

[0389] The segmentation unit 652a divides the multiple waveform signals input from the differentiation unit 651a into segmented waveform data with integer periods. In this embodiment, the segmentation unit 652a sets the integer period to one period, but it can also set it to multiple periods. The segmentation unit 652a outputs the segmented waveform data to the normalization unit 653a.

[0390] The normalization unit 653a performs normalization to unify the time width of the multiple segmented waveform signals input from the segmentation unit 652a, obtains an average waveform signal that is the average of the multiple segmented waveform signals, and performs normalization by setting the maximum value of the average waveform signal amplitude to 1 and the minimum value to 0, thereby obtaining evaluation data. The normalization unit 653a outputs the obtained evaluation data to the second mode selection unit 66a.

[0391] The normalization unit 653a can trim multiple segmented waveform signals input from the segmentation unit 652a with a certain time width or a certain number of samples to unify the time width of the segmented waveform signals. The method for unifying the time width by the normalization unit 653a is determined by the filtering processing unit 650a.

[0392] When the normalization unit 653a obtains the average waveform signal, it needs multiple segmented waveform signals. The number of segmented waveform signals required is determined by the filtering processing unit 650a.

[0393] The second mode selection unit 66a, referring to a pre-obtained second classification mode, classifies the evaluation data input from the normalization unit 653a into a second mode. The second mode selection unit 66a then outputs the classified second mode to the first processing selection unit 63a.

[0394] The first processing selection unit 63a refers to a pre-acquired processing mode and selects a first processing mode based on the first mode input from the first mode selection unit 62a and the second mode input from the second mode selection unit 66a. The first processing selection unit 63a outputs the selected first processing mode to the blood glucose value acquisition unit 64a.

[0395] The biological information obtained by the biological information processing system 100 includes blood glucose levels. Other examples include blood pressure, blood oxygen saturation, blood carbon dioxide concentration, vascular age, and the degree of diabetes. The biological information processing system 100 may also obtain other information related to pulse wave signals as biological information.

[0396] Next, an example of the operation of the biological information processing system 100 in this embodiment will be described. Figure 26 This is a flowchart illustrating an example of the operation of the biological information processing system 100 in this embodiment.

[0397] First, in the pulse wave signal acquisition step S10a, the acquisition unit 50 acquires the pulse wave signal and outputs it to the communication I / F 51 via the internal bus 54. In the pulse wave signal acquisition step S10a, for example, instead of the pulse wave signal acquired by the acquisition unit 50, the pulse wave signal recorded in the memory 52 may be output to the communication I / F 51 via the internal bus 54. For example, an FBG sensor may be used to measure the pulse wave signal.

[0398] Next, the communication I / F 51, which has received the pulse wave signal from the acquisition unit 50, sends the pulse wave signal to the pulse wave signal acquisition unit 60a via the communication network 3. Alternatively, at this time, the pulse wave signal stored in the server 4 can be sent to the pulse wave signal acquisition unit 60a instead of the pulse wave signal acquired by the acquisition unit 50.

[0399] Next, the pulse wave signal acquisition unit 60a, which has transmitted the pulse wave signal via the communication network 3, outputs the pulse wave signal to the classification data extraction unit 61a and the evaluation data extraction unit 65a.

[0400] The classification data extraction unit 61a extracts classification data from the pulse wave signal input from the pulse wave signal acquisition unit 60a, referring to extraction conditions. The classification data is waveform data extracted by the classification data extraction unit 61a for classifying the pulse wave signal by the first mode selection unit 62a. The classification data is, for example, waveform data obtained by processing the pulse wave signal into a waveform of one cycle through at least one of the following processes: filtering, differentiation, normalization, and averaging.

[0401] The filtering processing unit 610a, referring to the extraction conditions, determines the extraction method of the pulse wave signal by the classification data extraction unit 61a based on the state and additional information of the acquired pulse wave signal. The additional information represents user information, such as at least one of the following: in addition to information including the user's age, gender, medical history, lifestyle habits, health status, medication information, degree of arteriosclerosis, or genetic information, it also includes environmental information such as temperature, humidity, or acceleration measured by the accelerometer installed on sensor 5.

[0402] For example, the filtering processing unit 610a issues a command to cause the differentiating unit 611a to differentiate the pulse wave signal input from the filtering processing unit 610a.

[0403] First, the classification data extraction unit 61a outputs the pulse wave signal input from the pulse wave signal acquisition unit 60a to the filtering processing unit 610a.

[0404] Next, after filtering the pulse wave signal input from the pulse wave signal acquisition unit 60a, the filtering processing unit 610a outputs the pulse wave signal to the differentiating unit 611a.

[0405] Next, the filtering processing unit 610a determines whether to differentiate the pulse wave signal input from the filtering processing unit 610a. After processing, the differentiating unit 611a outputs the pulse wave signal to the dividing unit 612a.

[0406] The reason why the filtering processing unit 610a determines whether the differentiation unit 611a should differentiate the pulse wave signal is that the characteristics of the classification data obtained by differentiating or not differentiating the pulse wave signal will be different, and the first processing appropriate for the pulse wave signal is determined based on the obtained pulse wave signal.

[0407] Next, the segmentation unit 612a divides the multiple waveform signals input from the differentiation unit 611a into segmented waveform data for one cycle, and then averages them. Then, the segmentation unit 612a outputs the segmented waveform data to the normalization unit 613a.

[0408] The normalization unit 613a normalizes the horizontal axis to unify the time width of the multiple segmented waveform signals input from the segmentation unit 612a, obtaining an average waveform signal that is the average of the multiple segmented waveform signals. Then, it normalizes the vertical axis by setting the maximum value of the average waveform signal to 1 and the minimum value to 0, obtaining classification data. The normalization unit 613a then outputs the classification data to the first mode selection unit 62a.

[0409] The reason for normalizing the horizontal axis using the normalization unit 613a to unify the time width is that significant differences are observed at the end of the pulse wave; therefore, this portion is removed, and the main body of the pulse wave is used as the analysis object. Furthermore, the reason for normalizing the vertical axis by setting the maximum value of the average waveform signal to 1 and the minimum value to 0 is to average the deviation in pressure applied when the FBG sensor is installed at the measurement site, and the deviation in measurement data caused by positional shift of the FBG sensor during measurement. This suppresses noise caused by measurement deviations and improves the accuracy of the correlation between the pulse wave signal and the measured values ​​of biological information.

[0410] Next, in the first mode selection step S12a, referring to the first classification mode, the first mode selection unit 62a classifies the classification data input from the classification data extraction unit 61a to obtain the first mode. For example, similar to the embodiment described above, the classification mode uses... Figure 12 Such classification patterns of acceleration pulse waves.

[0411] When the first mode selection unit 62a receives classification data based on the acceleration pulse wave after differentiating the pulse wave signal by the differentiating unit 611a, it determines whether the classification data conforms to, for example... Figure 12 Which mode is selected determines the first mode.

[0412] Next, in the evaluation data extraction step S13a, the evaluation data extraction unit 65a extracts evaluation data from the pulse wave signal input from the pulse wave signal acquisition unit 60a, referring to the evaluation extraction conditions. The extraction method performed by the evaluation data extraction unit 65a on the pulse wave signal is largely the same as that performed by the classification data extraction unit 61a, but the difference is that in the classification data extraction unit 61a, extraction conditions are referenced, while in the evaluation data extraction unit 65a, evaluation extraction conditions are referenced instead of extraction conditions. Furthermore, a pulse wave signal different from the pulse wave signal input from the pulse wave signal acquisition unit 60a to the evaluation data extraction unit 65a can be input from the pulse wave signal acquisition unit 60a to the classification data extraction unit 61a, and the evaluation data extraction unit 65a extracts evaluation data from the aforementioned pulse wave signal. The evaluation data is waveform data extracted by the evaluation data extraction unit 65a for classifying the pulse wave signal by the second mode selection unit 66a or obtaining biological information by the blood glucose value acquisition unit 64a. The evaluation data is waveform data obtained, for example, by processing the pulse wave signal into a waveform of one cycle through at least one of the following processes: filtering, differentiation, normalization, and averaging.

[0413] The filtering processing unit 650a determines the extraction method for the pulse wave signal by the evaluation data extraction unit 65a based on the state and additional information of the obtained pulse wave signal, referring to the extraction conditions for evaluation. As an extraction method, for example, when the above-mentioned high acceleration value is measured as additional information, the filtering processing unit 650a can use a bandpass filter with a narrower bandwidth.

[0414] Furthermore, the filtering processing unit 650a determines the extraction method of the pulse wave signal performed by the evaluation data extraction unit 65a based on the extraction method of the classification data. For example, if the extraction method of differentiation in the differentiation unit 611a is determined, the filtering processing unit 650a issues a command so that the differentiation unit 651a does not differentiate the pulse wave signal input from the filtering processing unit 650a. As a result, easily classifiable acceleration pulse waves can be classified as classification data, a first mode corresponding to the classification data can be selected, a first processing corresponding to the first mode can be determined, and velocity pulse waves, which are easy to suppress false detections, can be used as evaluation data to obtain blood glucose values ​​corresponding to the evaluation data, enabling more accurate evaluation.

[0415] First, the evaluation data extraction unit 65a outputs the pulse wave signal input from the pulse wave signal acquisition unit 60a to the filtering processing unit 650a.

[0416] Next, after filtering the pulse wave signal input from the pulse wave signal acquisition unit 60a, the filtering processing unit 650a outputs the pulse wave signal to the differentiating unit 651a.

[0417] Next, the filtering processing unit 650a determines whether to differentiate the pulse wave signal input from the filtering processing unit 650a. After processing, the differentiating unit 651a outputs the pulse wave signal to the dividing unit 652a.

[0418] Next, the segmentation unit 652a divides the multiple waveform signals input from the differentiation unit 651a into segmented waveform data for one cycle, and then averages them. Then, the segmentation unit 652a outputs the segmented waveform data to the normalization unit 653a.

[0419] The normalization unit 653a normalizes the horizontal axis to unify the time width of the multiple segmented waveform signals input from the segmentation unit 652a, obtaining an average waveform signal that is the average of the multiple segmented waveform signals. Then, it normalizes the vertical axis by setting the maximum value of the average waveform signal to 1 and the minimum value to 0, obtaining evaluation data. Finally, the normalization unit 653a outputs the evaluation data to the second mode selection unit 66a.

[0420] The reason for normalizing the horizontal axis using the normalization unit 653a to unify the time width is that significant differences are observed at the end of the pulse wave; therefore, this portion is removed, and the main body of the pulse wave is used as the analysis object. Furthermore, the reason for normalizing the vertical axis by setting the maximum value of the average waveform signal to 1 and the minimum value to 0 is to average the deviation in pressure applied when the FBG sensor is installed at the measurement site, and the deviation in measurement data caused by positional shift of the FBG sensor during measurement. This suppresses noise caused by measurement deviations and improves the accuracy of the correlation between the pulse wave signal and the measured values ​​of biological information.

[0421] Next, in the second mode selection step S14a, referring to the second classification mode, the second mode selection unit 66a classifies the evaluation data input from the evaluation data extraction unit 65a to obtain the second mode.

[0422] Similar to the first mode selection unit 62a, the second mode selection unit 66a determines which mode the input evaluation data and additional information match, and decides on the second mode.

[0423] In addition, similar to the first mode selection unit 62a, the second mode selection unit 66a can also weight each classification result based on the input multiple classification data and additional information, and set the ratio of the multiple modes to the second mode.

[0424] Next, in the first processing selection step S15a, the first processing selection unit 63a refers to the pre-acquired processing mode, selects the first processing corresponding to the first mode input from the first mode selection unit 62a and the second mode input from the second mode selection unit 66a, and outputs the first processing to the blood glucose value acquisition unit 64a.

[0425] The processing mode is a data set that includes a series of processing methods performed on the evaluation data by the blood glucose value acquisition unit 64a in order to obtain blood glucose values ​​based on the evaluation data. The first processing selection unit 63a determines the processing method performed on the evaluation data as the first processing from the processing mode. The data set of the processing mode may also include multiple processing methods. As a processing mode, for example, a processing method that uses a calibration model representing the correlation between the measured value and the pulse wave signal to obtain blood glucose values ​​based on the evaluation data can be cited. In addition, the processing mode includes, for example, a processing method that estimates outliers in the blood glucose value based on the deviation between the input evaluation data and the calibration model.

[0426] The first processing selection unit 63a may also select the processing method that uses the calibration model most suitable for the first mode input from the first mode selection unit 62a and the second mode input from the second mode selection unit 66a as the first processing method, and output the first processing method to the blood glucose value acquisition unit 64a.

[0427] As a method for selecting the first processing, for example, a processing mode using a calibration model may be selected as the first processing, and this first processing may be output to the blood glucose value acquisition unit 64a. This calibration model is constructed by performing regression analysis using waveform data of the pulse wave pattern that best matches the first mode input from the first mode selection unit 62a and the second mode input from the second mode selection unit 66a as explanatory variables, and using the measured blood glucose value as the target variable, and based on the analysis results. For example, if mode B is input as the first mode, a processing method using a calibration model may be selected as the first processing mode. This calibration model is constructed by performing regression analysis using waveform data of mode B as explanatory variables and the measured blood glucose value as the target variable, and based on the analysis results.

[0428] Furthermore, when the first mode input from the first mode selection unit 62a and the second mode input from the second mode selection unit 66a consist of multiple modes, the first processing selection unit 63a may also select multiple first processes that use multiple calibration models that correspond to each mode, and output the multiple first processes to the blood glucose value acquisition unit 64a.

[0429] Furthermore, the first processing selection unit 63a can also select a first processing method corresponding to the input additional information. For example, if the user's age of over 40 is input as additional information, the first processing selection unit 63a can select a processing method that uses a calibration model representing the correlation between measured values ​​from a user in their 40s and the pulse wave signal as the first processing method, and output this first processing method to the blood glucose value acquisition unit 64a. Alternatively, for example, if the degree of arteriosclerosis, the presence of diabetes, etc., are used as additional information, as described above, the first processing selection unit 63a can select a processing method that uses a calibration model representing the correlation between measured values ​​from a user matching the additional information and the pulse wave signal as the first processing method, and output this first processing method to the blood glucose value acquisition unit 64a.

[0430] Next, in the blood glucose value acquisition step S16a, the blood glucose value acquisition unit 64a refers to the first process input from the first process selection unit 63a and acquires biological information such as blood glucose value corresponding to the evaluation data input from the evaluation data extraction unit 65a.

[0431] When multiple first processes are input from the first process selection unit 63a, the blood glucose value acquisition unit 64a can also acquire multiple blood glucose values ​​corresponding to the evaluation data based on each first process. Furthermore, in this case, for example, an optimal blood glucose value can be calculated based on the multiple blood glucose values. As a method for calculating the optimal blood glucose value, the average of the multiple blood glucose values ​​can be output as the optimal blood glucose value. Alternatively, for example, each blood glucose value can be weighted according to its respective measurement accuracy, and the optimal blood glucose value can be calculated based on the weighting of the multiple blood glucose values. Other examples include: outputting the blood glucose value obtained by the first process that shows a good value on the error grid among the multiple first processes that have acquired blood glucose values ​​as the optimal blood glucose value; acquiring multiple blood glucose values ​​using each first process, and outputting the blood glucose value obtained by the first process with the smallest deviation as the optimal blood glucose value; evaluating whether the blood glucose value is within a predetermined allowable range, and outputting the blood glucose value that is within the range as the optimal blood glucose value.

[0432] Regarding the pulse wave waveform signal, the blood glucose value obtained from the pulse wave signal can sometimes deviate depending on the user's gender, age, and other attributes. For example, when comparing the pulse waves measured from a man in his 20s with those measured from a woman in her 50s, the accuracy of the obtained blood glucose value may vary depending on the processing method, and may not be sufficiently accurate.

[0433] In contrast, the bio-information processing system 100 in this embodiment includes a pulse wave signal acquisition step S10a, which acquires a velocity pulse wave as a pulse wave signal; a classification data extraction step S11a, which extracts classification data based on the pulse wave signal; an evaluation data extraction step S13a, which extracts evaluation data based on the pulse wave signal using an extraction method different from that used for the classification data, referring to evaluation extraction conditions; a first mode selection step S12a, which selects one or more first modes corresponding to the classification data, referring to a first classification mode that includes multiple pre-acquired first modes; a first processing selection step S15a, which selects one or more first processes corresponding to the first mode, referring to a processing mode that includes multiple pre-acquired first processes; and a bio-information acquisition step (e.g., blood glucose value acquisition step S16a) which acquires bio-information such as blood glucose value based on the evaluation data according to the first process.

[0434] That is, according to this embodiment, the bio-information processing system 100 refers to the first processing corresponding to the first mode selected in the first processing selection step S15a, and obtains bio-information such as blood glucose levels based on the evaluation data extracted in the evaluation data extraction step S13a. Therefore, it is possible to select the most suitable processing method for the input pulse wave signal and obtain bio-information such as blood glucose levels based on the input pulse wave signal. Thus, it is possible to perform pulse wave signal processing that matches the user's attributes and obtain highly accurate evaluation results.

[0435] Furthermore, according to this embodiment, the bio-information processing system 100 differentiates the pulse wave signal in the classification data extraction step S11a, thereby extracting the acceleration pulse wave as classification data and selecting the first mode corresponding to the classification data. In contrast, in the evaluation data extraction step S13a, the pulse wave signal is not differentiated, thereby extracting the velocity pulse wave as evaluation data. Therefore, by classifying the characteristics of the pulse wave signal using the acceleration pulse wave suitable for classification, and by using the velocity pulse wave capable of suppressing false detections to measure blood glucose levels, high-precision evaluation is possible.

[0436] Furthermore, according to this embodiment, in the first processing selection step S15a, a first processing method corresponding to the first mode selected in the first mode selection step S12a and the second mode selected in the second mode selection step S14a is selected. Therefore, a more suitable processing method can be selected based on multiple modes obtained from a single pulse wave signal that employ different extraction methods, further improving accuracy.

[0437] Furthermore, according to this embodiment, in the first mode selection step S12a, a first mode corresponding to the classification data and additional information extracted through the classification data extraction step S11a is selected. Therefore, the first mode corresponding to the additional information can be selected, further improving accuracy.

[0438] Furthermore, according to this embodiment, in the first processing selection step S15a, a first processing corresponding to the first pattern and additional information extracted through the classification data extraction step S11a is selected. Therefore, the first processing corresponding to the additional information can be selected, further improving accuracy.

[0439] Embodiments of the present invention have been described, but these embodiments are provided as examples and are not intended to limit the scope of the invention. This novel embodiment can be implemented in various other ways, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. This embodiment, and its variations, are included within the scope and spirit of the invention, and are included within the scope of the invention as described in the claims and its equivalents.

[0440] Label Explanation

[0441] 1: Biological information processing device

[0442] 3: Communication Network

[0443] 4: Server

[0444] 5: Sensors

[0445] 6: Testing Department

[0446] 10: Shell

[0447] 11: Obtaining Department

[0448] 12: Generation Department

[0449] 13: Output Department

[0450] 14: Storage Department

[0451] 15: Academic Department

[0452] 50: Obtaining Department

[0453] 51: Communication I / F

[0454] 52: Memory

[0455] 53: Command Department

[0456] 54: Internal Bus

[0457] 55: Wristband

[0458] 100: Biological Information Processing System

[0459] 101: CPU

[0460] 102: ROM

[0461] 103: RAM

[0462] 104: Preservation Department

[0463] 105: I / F

[0464] 106: I / F

[0465] 107: I / F

[0466] 108: Input Section

[0467] 109: Display Section

[0468] 110: Internal Bus

[0469] S110: Obtaining Steps

[0470] S120: Generation Steps

[0471] S130: Output Steps

[0472] S140: Saving Steps

[0473] S150: Comprehensive Evaluation Steps

[0474] S160: Calculation Steps

[0475] S170: Update Steps

Claims

1. A biological information processing system that evaluates a user's biological information. The biological information processing system is characterized by having: The acquisition unit performs different types of processing on one pulse wave data corresponding to either the velocity pulse wave or the acceleration pulse wave based on the user's pulse wave, which is suitable for the biological information to be estimated, thereby obtaining first evaluation data and second evaluation data, wherein the first evaluation data and the second evaluation data are data corresponding to either the velocity pulse wave or the acceleration pulse wave of the user. A database storing classification information, which is generated by using multiple learning data sets as a pair of input data and reference data, wherein... The input data is based on pre-acquired training pulse wave data, and the reference data is data containing biological information associated with the input data; and The generation unit, referring to the database, generates a first evaluation result containing first organism information corresponding to the first evaluation data. This first evaluation result is health information obtained based on the first organism information. The generation unit generates a second evaluation result containing information about a second organism of a different type than the information about the first organism, which corresponds to the second evaluation data. The classification information includes first classification information and second classification information generated using different types of the learning data. The generation unit refers to the first classification information and generates the first evaluation result corresponding to the first evaluation data. The generation unit refers to the second classification information and generates the second evaluation result corresponding to the second evaluation data.

2. The biological information processing system according to claim 1, characterized in that, The biological information processing system also has a storage unit, which stores the first evaluation result and the second evaluation result.

3. The biological information processing system according to claim 1, characterized in that, The biological information processing system has a comprehensive evaluation unit. The comprehensive evaluation unit obtains additional information representing the characteristics of the user, and generates a comprehensive evaluation result based on the first evaluation result, the second evaluation result, and the additional information, which comprehensively evaluates the characteristics of the user.

4. The biological information processing system according to claim 1, characterized in that, The classification information includes multiple attribute-based classification information calculated using different learning data. The generation unit includes: The selection unit, referring to the second evaluation data, selects the first classification information from a plurality of attribute-based classification information; and The attribute generation unit generates the first evaluation result corresponding to the first evaluation data by referring to the first classification information.

5. The biological information processing system according to claim 4, characterized in that, The acquisition unit acquires data corresponding to the velocity pulse wave based on the pulse wave as the first evaluation data. The acquisition unit acquires data corresponding to the acceleration pulse wave based on the pulse wave as the second evaluation data.

6. The biological information processing system according to claim 1, characterized in that, Based on the characteristics of the pulse wave, the generation unit selects the first category information from the classification information. The generation unit refers to the first classification information and generates the first evaluation result corresponding to the first evaluation data.

7. The biological information processing system according to claim 2, characterized in that, The biological information processing system has a computing unit, which generates a comprehensive evaluation result that comprehensively evaluates the user's characteristics based on the first evaluation result and the second evaluation result saved by the storage unit.

8. The biological information processing system according to claim 2, characterized in that, The storage unit obtains the judgment result of the user's evaluation of the first evaluation result and the second evaluation result. The storage unit associates and stores the judgment result, the first evaluation result, and the second evaluation result respectively.

9. The biological information processing system according to claim 8, characterized in that, The biological information processing system has an update unit, which updates the classification information based on the judgment result, the first evaluation result, and the second evaluation result stored in the storage unit.

10. The biological information processing system according to claim 1, characterized in that, The biological information processing system has the following features: A server, which stores the database, and generates the first evaluation result through the generation unit; and A biological information processing device that receives and displays the first evaluation result from the server.

11. The biological information processing system according to claim 3, characterized in that, The biological information processing system also has a storage unit for storing a data structure containing the first evaluation result and the second evaluation result.

12. A server, characterized in that, The server stores the first evaluation result as described in claim 1.

Images