JP2022187119A - Information processing device, blood pressure estimation method, and program - Google Patents

Abstract

To provide an information processing device capable of estimating a blood pressure in consideration of a plurality of factors which have influences on the blood pressure.SOLUTION: The information processing device comprises: a biological information acquisition part which acquires biological information of a measured person; a feature amount determination part which determines a plurality of feature amounts from the biological information; a class selection part which selects a belonging class of the measured person, from a plurality of classes classified with a reference related to a first feature amount included in the plurality of feature amounts; and a blood pressure estimation part which outputs an estimated blood pressure by inputting a second feature amount included in the feature amounts to a blood pressure estimation model selected in accordance with the class.SELECTED DRAWING: Figure 2

Description

The present disclosure relates to an information processing device, a blood pressure estimation method, and a program.

In Patent Document 1, the pulse wave transit time and heart rate at rest are measured in advance as reference values, and the pulse wave transit time and heart rate fluctuations from the subject's reference values are used as the basis for measurement. , discloses a technique for correcting parameter coefficients for calculating blood pressure.

Japanese Patent Application Laid-Open No. 09-220207

When the blood pressure is measured by compressing the blood vessel, the subject feels uncomfortable. However, since blood pressure fluctuates due to multiple factors, it is necessary to consider multiple factors that affect blood pressure in order to estimate blood pressure without compressing blood vessels. In the technique disclosed in Patent Literature 1, in order to estimate the blood pressure without compressing the blood vessels, it is necessary to measure in advance the pulse wave transit time and the heart rate at rest as reference values. Therefore, the technique disclosed in Patent Document 1 may not be able to accurately calculate the blood pressure of a subject whose resting pulse wave transit time and heart rate have not been measured. Accordingly, an object of one aspect of the present disclosure is to provide an information processing apparatus, a blood pressure estimation method, and a program capable of estimating blood pressure in consideration of a plurality of factors affecting blood pressure.

An information processing apparatus according to an aspect of the present disclosure includes a biological information acquisition unit that acquires biological information of a subject, a feature amount determination unit that determines a plurality of feature amounts from the biological information, and a classification selection unit that selects a classification to which the subject belongs from a plurality of classifications classified according to a criterion related to the included first feature value; and a blood pressure estimating unit that inputs a second feature quantity included in the feature quantity of and outputs an estimated blood pressure.

A blood pressure estimation method according to an aspect of the present disclosure includes a step of acquiring biometric information of a subject, a step of determining a plurality of feature amounts from the biometric information, and a first feature included in the plurality of feature amounts. selecting a category to which the person to be measured belongs from a plurality of categories classified according to criteria related to quantities; and inputting the feature amount of and outputting the estimated blood pressure.

A program according to an aspect of the present disclosure provides a computer with a function of acquiring biometric information of a subject, a function of determining a plurality of feature amounts from the biometric information, and a first feature included in the plurality of feature amounts. a function of selecting a classification to which the subject belongs from a plurality of classifications classified according to criteria related to feature quantities; and a function of inputting the feature quantity of 2 and outputting an estimated blood pressure.

It is a figure which shows an example of a measurement system. It is a figure which shows an example of a structure of the information processing apparatus which concerns on 1st embodiment. It is a figure which shows an example of the classification of the feature-value contained in a feature-value set. It is a figure which shows an example of the classification of the feature-value contained in a feature-value set. It is a figure which shows an example of the classification of the feature-value contained in a feature-value set. It is a figure which shows an example of reference|standard information. It is a figure which shows an example of a classification. 4 is a flowchart showing an example of processing executed by the information processing apparatus according to the first embodiment; 6 is a flowchart showing an example of processing subsequent to FIG. 5; FIG. 7 is a flowchart showing an example of processing subsequent to FIG. 6; FIG. It is a figure which shows the feature-value contained in the 2nd feature-value corresponding to each classification. It is a figure which shows an example of the coefficient and intercept of the blood-pressure estimation formula corresponding to each classification. It is a figure which shows an example of the coefficient of a blood-pressure estimation formula, and an intercept which concerns on the modification of 1st embodiment. It is a figure which shows an example of a structure of the information processing apparatus which concerns on 2nd embodiment. 9 is a flowchart showing an example of processing for selecting a category by the information processing apparatus according to the second embodiment;

(First embodiment)
A first embodiment will be described with reference to FIGS. 1 to 9. FIG. In the drawings, the same or equivalent elements are denoted by the same reference numerals, and overlapping descriptions are omitted.

The configuration of the information processing device 103 will be described with reference to FIGS. 1 to 4B. FIG. 1 is a diagram showing an example of a measurement system 100. As shown in FIG. The measurement system 100 includes an imaging device 102 and an information processing device 103 .

The imaging device 102 photographs the subject 101 facing the imaging device 102 and acquires an image including the image of the body surface of the subject 101 . Specifically, the imaging device 102 captures images of the person to be measured 101 at a predetermined frame rate for a predetermined period of time to obtain an image including an image of the body surface of the measurement target 110 . For example, the imaging device 102 captures the subject 101 at 15 to 300 fps (frames per second) for 0.1 to 120 seconds to obtain an image including the body surface of the subject 101 . For example, imaging device 102 is an RGB camera.

The information processing device 103 estimates the blood pressure of the person to be measured 101 based on the images and the like acquired by the imaging device 102 . For example, the information processing device 103 is a PC (Personal Computer), a smart phone, a tablet terminal, a terminal dedicated to blood pressure estimation, or the like.

FIG. 2 is a diagram showing an example of the configuration of the information processing device 103. As shown in FIG. The information processing device 103 includes a storage unit 201 and a control unit 202 .

The storage unit 201 is a recording medium capable of recording various data, programs, etc., and is configured by, for example, a hard disk, an SSD (Solid State Drive), a semiconductor memory, or the like. Storage unit 201 stores blood pressure estimation model database 220 and reference information 221 . The blood pressure estimation model database 220 stores multiple blood pressure estimation models 222 . The blood pressure estimation model 222 indicates the relationship between the feature amount related to the living body and the blood pressure of the living body.

The reference information 221 indicates criteria for classifying a plurality of classifications related to biometric features. When the criterion information 221 indicates a plurality of criteria, the plurality of criteria indicate criteria related to different types of information. Specifically, the criteria indicated by the criteria information 221 include criteria regarding information related to the physiological state of the living body. For example, the criteria indicated by the criteria information 221 include at least one selected from the group consisting of blood pressure criteria and pulse rate criteria. Blood pressure standards are standards related to systolic blood pressure, diastolic blood pressure, pulse pressure, mean blood pressure, and the like. Also, for example, the criteria indicated by the criteria information 221 may include criteria relating to information other than blood pressure and pulse rate.

Also, for example, the criteria indicated by the criteria information 221 may include criteria regarding information related to attributes of a living body. For example, criteria related to information related to attributes of a living body are criteria related to sex, age, height, weight, and the like.

The control unit 202 executes various processes according to programs and data stored in the storage unit 201 . The control unit 202 is realized by a processor such as a CPU (Central Processing Unit), for example.

The control unit 202 includes a light source adjustment unit 203, a body surface area extraction unit 204, a pulse wave calculation unit 205, a biological information acquisition unit 206, a feature amount determination unit 207, a classification selection unit 208, and a feature amount selection unit. 209 and a blood pressure estimation unit 210 .

The light source adjustment unit 203 adjusts the intensity of the light source in the imaging environment in which the imaging device 102 images the person 101 to be measured.

A body surface region extraction unit 204 extracts a body surface region from the image 231 acquired by the imaging device 102 . The body surface area is an image area of the body surface of the subject 101 included in the image 231 and includes a plurality of pixels. The body surface area may be an area obtained by dividing the image of the subject's 101 face area. For example, the body surface region may be the forehead region, cheek region, fingertip region, wrist region, palm region, and the like of the subject 101 . Alternatively, the body surface area may be an area obtained by dividing the forehead area, cheek area, fingertip area, wrist area, palm area, or the like of the subject 101 .

The pulse wave calculator 205 calculates a pulse wave 232 from time-series signals indicating pixel values of the body surface region corresponding to the same position on the body surface. In the present disclosure, a pulse wave 232 is a time-series signal that indicates changes in blood vessel volume with respect to the same position on the body surface.

The biometric information acquisition unit 206 acquires the biometric information 233 of the subject 101 . Biological information 233 may include physiological state information 234 . The physiological state information 234 indicates the physiological state of the subject 101 . The physiological state information 234 may include information indicating an electrocardiographic signal, body surface temperature, and the like. Furthermore, the biometric information 233 includes attribute information 235 . The attribute information 235 indicates attributes of the subject 101 . For example, the attribute information 235 indicates the age, sex, height, weight, etc. of the subject 101 .

The feature amount determination unit 207 determines a plurality of feature amounts from the biometric information 233 . Specifically, the feature amount determination unit 207 determines a feature amount set 236 representing a plurality of feature amounts from the biometric information 233 . A plurality of feature amounts indicated by the feature amount set 236 indicate different types of features. It is assumed that the type of each feature amount is determined in advance. A plurality of feature amounts indicated by the feature amount set 236 include feature amounts related to at least one selected from the group consisting of blood pressure and pulse rate.

Classification selection section 208 selects classification 238 to which subject 101 belongs from a plurality of classifications. The multiple classifications are classified by criteria associated with a first feature 237 included in the multiple features indicated by feature set 236 . The criteria indicated by criteria information 221 relate to first feature quantity 237 .

The feature amount selection unit 209 selects a second feature amount 239 from a plurality of feature amounts indicated by the feature amount set 236 . Blood pressure estimator 210 selects blood pressure estimation model 222 from blood pressure estimation model database 220 according to classification 238 . Specifically, blood pressure estimation section 210 inputs second feature quantity 239 to blood pressure estimation model 222 selected according to classification 238 and outputs estimated blood pressure 240 . The estimated blood pressure 240 is an estimated blood pressure of the subject 101 .

3A to 3C are diagrams showing examples of the feature amounts _x1 to _xn included in the feature amount set 236. FIG. n is a natural number. x _k indicates each feature included in the feature set 236 . Here, k is a natural number greater than or equal to 1 and less than or equal to n.

For example, the feature set 236 illustrated in FIG. 3A includes features _x1 through _x4 . In the feature quantity set 236 illustrated in FIG. 3A, the feature quantities x1 to x4 _respectively indicate the time of _one pulse wave, the rising angle of the pulse wave, the integrated value of one pulse wave, and the body weight. Also, for example, the feature quantity set 236 illustrated in FIG. 3B includes the feature quantity _x1 to the feature quantity _x5 . In the feature quantity set 236 illustrated in FIG. 3B, the feature quantity x1 to the feature quantity _x5 indicate the rising angle of the pulse wave, the R _- wave interval of the electrocardiographic signal, the body surface temperature, the age, and the height, respectively. Also, for example, the feature amount set 236 illustrated in FIG. 3C includes feature amounts _x1 to _x6 . In the feature quantity set 236 illustrated in FIG. 3C, the feature quantity x1 to feature quantity _x6 are _respectively the pulse wave peak interval, pulse wave rise angle, pulse wave integral value, height, weight, and gender. show. Note that the feature amounts included in the feature amount set 236 are not limited to the six types described above. For example, the feature amounts included in the feature amount set 236 may indicate pulse wave fall time, pulse wave amplitude, and the like.

FIG. 4A is a diagram showing an example of the reference information 221. As shown in FIG. FIG. 4B is a diagram showing an example of classifications 401 to 404 classified according to the criteria indicated by the criteria information 221 illustrated in FIG. 4A. Classification 401 to classification 404 are classified according to the criterion of whether the systolic blood pressure is 130 mmHg or more and the criterion of whether the pulse rate is 70 bpm or more. Specifically, the reference information 221 illustrated in FIG. 4A indicates that classification 401 is performed when the systolic blood pressure is 130 mmHg or more and the pulse rate is 70 bpm or more. Further, the reference information 221 illustrated in FIG. 4A indicates that classification 402 is performed when the systolic blood pressure is 130 mmHg or more and the pulse rate is less than 70 bpm. Further, the reference information 221 illustrated in FIG. 4A indicates that the systolic blood pressure is less than 130 mmHg and the pulse rate is less than 70 bpm are classified into the classification 403 . Further, the reference information 221 illustrated in FIG. 4A indicates that the systolic blood pressure is less than 130 mmHg and the pulse rate is 70 bpm or more is classified into the classification 403 .

5 to 7 are flowcharts showing an example of processing executed by the information processing apparatus 103. FIG. In this example, the control unit 202 starts the process of step S501 illustrated in FIG. 5 after the image capturing device 102 captures images of the person to be measured 101 for a predetermined time. When the imaging device 102 takes an image, the light source adjustment unit 203 adjusts the intensity of the light source in the imaging environment in which the imaging device 102 images the body surface of the subject 101 . However, if the shooting environment is a lighting environment in which the body surface of the subject 101 can be properly shot, the light source adjustment unit 203 does not have to adjust the intensity of the light source.

Furthermore, it is assumed that the reference information 221 is stored in the storage unit 201 at the time when the control unit 202 starts the processing of step S501 illustrated in FIG. Furthermore, it is assumed that a plurality of blood pressure estimation models 222 are stored in the blood pressure estimation model database 220 when the control unit 202 starts the process of step S501. For example, the information processing apparatus 103 learns using one or more techniques such as regression, pattern recognition, multivariate analysis, statistical analysis, and machine learning before starting the process of step S501, and obtains a blood pressure estimation model. 222 is generated. For example, the information processing apparatus 103 uses biological information including the physiological state information and attribute information of the person to be measured 101 as an explanatory variable, and the blood pressure measured by pressing the arm of the person to be measured 101 as an objective variable. Calculate the weighting factor of the regression equation of The information processing device 103 then learns the weighting factors and the like calculated for the plurality of subjects 101 to generate the blood pressure estimation model 222 . The blood pressure estimation model 222 includes a blood pressure estimation formula including coefficients corresponding to classifications among a plurality of classifications and variables to which the second feature amount is input. The information processing device 103 stores the generated blood pressure estimation model 222 in the storage unit 201 .

First, the processing executed by the information processing apparatus 103 will be described with reference to FIG.

In step S<b>501 , the body surface region extraction unit 204 acquires the image 231 acquired by the imaging device 102 and stores the acquired image 231 in the storage unit 201 . In step S<b>502 , the body surface region extraction unit 204 extracts a body surface region from the image 231 stored in the storage unit 201 .

In step S<b>503 , the pulse wave calculator 205 calculates the pulse wave 232 . Specifically, the pulse wave calculator 205 extracts pixel value information from the body surface region of the subject 101 . The pixel value information is a statistical value of pixel values of each pixel in the body surface region for each of R (Red), G (Green), and B (Blue). For example, the pixel value statistic is the mean, median or mode of pixel values. Then, the pulse wave calculation unit 205 extracts a signal indicating a temporal change in pixel value information from the moving image of the body surface region over a predetermined period of time. Then, the pulse wave calculator 205 processes the extracted signal by independent component analysis, pigment component separation, band-pass filter, etc., and calculates the processed result as a pulse wave 232 .

In step S<b>504 , the biological information acquisition unit 206 acquires physiological state information 234 . For example, the biological information acquisition unit 206 acquires physiological state information 234 indicating an electrocardiographic signal and body surface temperature.

In step S<b>505 , the biometric information acquisition unit 206 acquires the attribute information 235 . The attribute information 235 indicates the age, sex, height, weight, etc. of the subject 101 . For example, the subject 101 inputs the age, sex, height, weight, etc. of the subject 101 using a keyboard, touch panel, or the like. In this case, the biometric information acquiring unit 206 acquires the input age, sex, height, weight, etc. of the person to be measured 101 as the attribute information 235 .

For example, the information processing apparatus 103 may present the subject 101 with age ranges such as 0 to 19 years old, 20 to 59 years old, and 60 years old or older. In that case, subject 101 selects the range to which his/her age belongs from the presented ranges. Similarly, the information processing apparatus 103 may present the subject 101 with height and weight ranges. In that case, the person to be measured 101 selects the range to which his or her height and weight belong from the presented ranges.

In step S506, the biometric information acquisition unit 206 generates biometric information 233 including the physiological state information 234 acquired in step S503 and the attribute information 235 acquired in step S504. Then, the control unit 202 shifts the processing to step S601 illustrated in FIG.

Next, the processing executed by the information processing apparatus 103 will be continued to be described with reference to FIG.

In step S<b>601 , the feature amount determination unit 207 determines a feature amount set 236 representing a plurality of feature amounts from the biometric information 233 . Specifically, the feature amount determination unit 207 determines a plurality of predetermined types of feature amounts from the biometric information 233, and generates a feature amount set 236 indicating the determined plurality of types of feature amounts.

A plurality of features represented by feature set 236 includes features determined from physiological state information 234 . For example, the feature amount determined from the physiological state information 234 indicates the characteristics of the pulse wave 232, and includes the integral value for one pulse wave, the required time for one pulse wave, the rising angle of the pulse wave 232, the pulse wave 232 , the amplitude of the pulse wave 232, and the like. The integrated value for one pulse wave shown as a feature amount may be the average value or the median value of the integrated values for one pulse wave for two or more pulse waves. The time required for one pulse wave indicated as a feature amount may be an average value or median value of the time required for one pulse wave for two or more pulse waves. Also, the rising angle of the pulse wave 232 indicated as a feature amount may be an average value or a median value of the rising angles for one pulse wave in two or more pulse waves. Similarly, the fall time of the pulse wave 232 and the amplitude of the pulse wave 232, which are indicated as feature amounts, are also average values or median values of the fall time or amplitude of one pulse wave in two or more pulse waves. may

Also, the plurality of feature amounts indicated by the feature amount set 236 may include feature amounts determined from the attribute information 235 . For example, when the attribute information 235 indicates gender, numerical values such as male=1 and female=2 are determined as feature amounts indicating the sex indicated by the attribute information 235 .

Further, for example, when the attribute information 235 indicates age, the feature amount determination unit 207 sets 0 to 19 years old=1, 20 to 59 years old=2, 60 as the feature amount indicating age indicated by the attribute information 235. Determine a numerical value such as age or older=3. Similarly, for example, when the attribute information 235 indicates height, weight, etc., the feature amount determination unit 207 determines numerical values for each range of height and weight as feature amounts indicating height, weight, etc. indicated by the attribute information 235. do.

In step S<b>602 , the classification selection unit 208 determines the first feature amount 237 related to the criterion indicated by the criterion information 221 from the plurality of feature amounts indicated by the feature amount set 236 . For example, the first feature quantity 237 may include a feature quantity related to at least one selected from the group consisting of blood pressure criteria and pulse rate criteria.

For example, when the reference indicated by the reference information 221 includes the blood pressure reference, the classification selection unit 208 determines the rising angle of the pulse wave 232 indicated by the feature amount set 236 as the feature amount included in the first feature amount 237. do. Also, for example, when the criteria indicated by the reference information 221 include the pulse rate criteria, the classification selection unit 208 assigns the duration of one beat of the pulse wave 232 indicated by the feature quantity set 236 to the first feature quantity 237. It is determined as a feature amount to be included.

In step S<b>603 , the classification selection unit 208 selects criteria to be determined from the criteria information 221 . In step S<b>604 , the classification selection unit 208 determines whether or not the feature amount related to the selected determination target criterion among the first feature amounts 237 satisfies the determination target criterion.

For example, suppose the criteria to be determined include a systolic blood pressure threshold. In that case, the classification selection unit 208 determines whether the systolic blood pressure calculated based on the rising angle of the pulse wave 232 indicated by the first feature amount 237 exceeds the threshold indicated by the criterion to be determined. .

Further, for example, it is assumed that the threshold value of the pulse rate to be determined is included. In that case, the classification selection unit 208 determines whether or not the pulse rate calculated based on the duration of one beat of the pulse wave 232 indicated by the first feature value 237 exceeds the threshold indicated by the criteria to be determined. judge.

In step S605, the classification selection unit 208 determines whether all the criteria indicated by the criteria information 221 have been determined. If it has not been determined whether or not the first feature amount 237 satisfies the determination target criteria for all the criteria indicated by the criteria information 221, the control unit 202 returns the process to step S603. In other words, the classification selection unit 208 repeats the processing from step S603 to step S605 until all criteria indicated by the criteria information 221 are determined. On the other hand, if it is determined whether or not the first feature amount 237 satisfies the determination target criteria for all the criteria indicated by the criteria information 221, the control unit 202 shifts the process to step S606.

In step S<b>606 , the classification selection unit 208 determines the classification 238 to which the subject 101 belongs from the judgment results for all criteria indicated by the criteria information 221 . Specifically, the classification selection unit 208 selects the classification 238 based on the relationship between the first feature quantity 237 and the criteria indicated by the criteria information 221 . Control unit 202 selects category 238 to which subject 101 belongs from a plurality of categories, based on first feature amount 237 and the reference indicated by reference information 221, by executing the processes from step S603 to step S606. to select. Then, the control unit 202 shifts the processing to step S701 illustrated in FIG.

Next, the processing executed by the information processing apparatus 103 will be continued with reference to FIG. 7 .

In step S<b>701 , the feature amount selection unit 209 selects the second feature amount 239 from a plurality of feature amounts indicated by the feature amount set 236 . In this example, the second feature quantity 239 is the feature quantity selected according to the classification 238 . The second feature quantity 239 may be the same as the first feature quantity 237 .

In step S<b>702 , the blood pressure estimation unit 210 selects the blood pressure estimation model 222 from the blood pressure estimation model database 220 according to the classification 238 . In step S<b>703 , the blood pressure estimation unit 210 inputs the second feature quantity 239 to the selected blood pressure estimation model 222 to calculate the estimated blood pressure 240 . The blood pressure estimation model 222 includes a blood pressure estimation formula including coefficients selected according to the classification 238 and variables to which the second feature quantity 239 is input. The blood pressure estimation unit 210 outputs the estimated blood pressure 240 in step S704. For example, the blood pressure estimator 210 displays characters indicating the estimated blood pressure 240 on a liquid crystal display or the like. Alternatively, for example, the blood pressure estimator 210 may output a voice indicating the estimated blood pressure 240 from a speaker.

For example, the blood pressure estimation unit 210 adds the coefficient W selected according to the classification 238 and the constant C according to the classification 238 to the blood pressure estimation formula Y=W·X+C of the subject 101 according to the classification 238. The estimated blood pressure Y of the person to be measured 101 is estimated by inputting a plurality of types of second feature quantities X. FIG. In other words, the blood pressure estimating unit 210 combines different types of feature quantities with weights according to the classification to which the person to be measured 101 belongs, which is selected based on the physiological condition and physical characteristics of the person to be measured 101. An estimated blood pressure of the subject 101 is estimated. Accordingly, the blood pressure estimation unit 210 estimates the blood pressure of the person to be measured 101 in consideration of a plurality of factors affecting the blood pressure of the person to be measured 101 according to the physiological condition and physical characteristics of the person to be measured 101. can be estimated.

FIG. 8A shows feature quantities included in the second feature quantities Xa ⁴⁰¹ to Xa ⁴⁰⁴ corresponding to the categories 401 to 404, respectively. A second feature amount Xa ⁴⁰¹ to a second feature amount Xa ⁴⁰⁴ indicate feature amounts included in the second feature amount 239 corresponding to the classifications 401 to 404, respectively. n1 to n4 illustrated in FIG. 8A indicate the number of feature quantities of the second feature quantities Xa ⁴⁰¹ to Xa ⁴⁰⁴ , respectively. n1 to n4 are natural numbers equal to or greater than 1 and equal to or less than the number of feature quantities indicated by the feature quantity set 236 . n1 to n4 may each have different values or may include the same value. The second feature quantities Xa ⁴⁰¹ to Xa ⁴⁰⁴ corresponding to the classifications 401 to 404 may each contain different numbers of feature quantities, or may contain the same number of feature quantities. Note that the categories 401 to 404 are examples, and the number of categories is not limited to four.

Next, referring to FIG. 8B, the blood pressure estimation formula of Ya=W ^m ·Xa ^m +C ^m indicated by the blood pressure estimation model 222 will be described as an example. Ya is the estimated blood pressure. Xa ^m indicates the second feature quantity 239 corresponding to the classification m, and is a vector of dimensions corresponding to the classification m. The coefficient W ^m is a vector of dimensions according to the classification m. Also, the intercept Cm is a constant corresponding to the classification ^m .

FIG. 8B shows an example of the coefficient W ^m and the intercept C ^m of the blood pressure estimation formula indicated by the blood pressure estimation model 222 corresponding to each of the classifications 401 to 404. FIG. A coefficient W ^m _k illustrated in FIG. 8B indicates a weight for the feature amount x ^m _k included in the second feature amount 239 corresponding to the classification m. In this example, m indicates one of categories 401 through 404 . For example, coefficient W ⁴⁰¹ illustrated in FIG. 8B includes coefficient w ⁴⁰¹ ₁ through coefficient w ⁴⁰¹ _n1 . Coefficients w ⁴⁰¹ ₁ to w ⁴⁰¹ _n1 respectively indicate weights for the respective feature amounts among the feature amounts x ⁴⁰¹ ₁ to x ⁴⁰¹ _n1 illustrated in FIG. 8A.

For example, when the feature quantity x ⁴⁰¹ ₁ indicates the duration of one pulse wave, the coefficient w ⁴⁰¹ ₁ indicates the weight for the duration of one pulse wave indicated by the feature quantity x ⁴⁰¹ ₁ . Also, for example, when the feature quantity x ⁴⁰¹ ₂ indicates the rising angle of the pulse wave, the coefficient w ⁴⁰¹ ₂ indicates the weight for the rising angle of the pulse wave indicated by the feature quantity x ⁴⁰¹ ₂ . Also, for example, when the feature quantity x ⁴⁰¹ ₃ indicates the integral value of one pulse wave, the coefficient w ⁴⁰¹ ₃ indicates the weight of the pulse wave indicated by the feature quantity x ⁴⁰¹ ₃ for the integral value of one pulse wave. Also, for example, when the feature quantity x ⁴⁰¹ ₄ indicates weight, the coefficient w ⁴⁰¹ ₄ indicates the weight for the weight indicated by the feature quantity x ⁴⁰¹ ₄ .

The physiological condition, physical characteristics, etc. of the person to be measured 101 affect the blood pressure of the person to be measured 101 . Therefore, when estimating the blood pressure without compressing the blood vessels, if the blood pressure is estimated using a single blood pressure estimation formula, there is a possibility that the blood pressure cannot be estimated accurately due to the influence of physiological conditions, physical characteristics, and the like. However, the information processing apparatus 103 according to the present embodiment can estimate the blood pressure by considering multiple factors that affect the blood pressure according to the physiological condition and physical characteristics of the subject 101. Become.

(Modification 1)
As a modified example 1 of the measurement system 100 according to the present embodiment, the imaging device 102 of the measurement system 100 may include a Doppler sensor instead of the RGB camera. When the imaging device 102 includes a Doppler sensor, the imaging device 102 detects the frequency of the emitted wave emitted from the Doppler sensor toward the subject 101 and the frequency of the reflected wave reflected by the body surface of the subject 101. is measured as the vibration of the body surface of the subject 101 . The vibration of the body surface measured by the Doppler sensor includes the vibration of the body surface due to changes in the blood pressure and volume of peripheral blood vessels accompanying heart beat. The biological information acquisition unit 206 extracts, from the signals measured by the Doppler sensor, signals indicating vibrations of the body surface caused by changes in the blood pressure and volume of peripheral blood vessels accompanying heart beats. Then, the biological information acquisition unit 206 acquires the signal indicating the vibration of the body surface of the subject 101 as the physiological state information 234 .

(Modification 2)
As a modified example 2 of the measurement system 100 according to the present embodiment, the blood pressure estimation formula for outputting the estimated blood pressure Yb may be the formula ^Yb = ^Wm ·Xb+Cm indicated by the blood pressure estimation model 222 . That is, the second feature quantity Xb may be the same regardless of the classification m.

For example, the second feature Xb may be identical to the plurality of features indicated by feature set 236 . In other words, the second feature Xb may be the same as all features indicated by the feature set 236 . Alternatively, the second feature quantity Xb may be the same as the first feature quantity 237 .

FIG. 9 is a diagram showing an example of the coefficient ^Wm and the intercept ^Cm of the blood pressure estimation formula according to this modification. The coefficient W ^m in this modified example includes the same number n of coefficients in the categories 401 to 404 . That is, the coefficient W ^m in this modification is a vector of the same dimension in the classifications 401 to 404 .

In this modified example, the coefficient w _k =0 is set for the feature quantity x _k that is not used for estimating the blood pressure among the feature quantities included in the second feature quantity Xb. k is a natural number equal to or greater than 1 and equal to or less than n. By setting coefficient w _k =0, blood pressure estimating section 210 can calculate estimated blood pressure 240 by excluding feature quantity x _k among second feature quantities Xb. Furthermore, since the coefficient W ^m including the coefficient w _k =0 differs depending on the classification m, the blood pressure estimation unit 210 calculates the estimated blood pressure 240 using the blood pressure estimation formula Yb corresponding to the selected classification 238. can.

(Modification 3)
As a third modification of the measurement system 100 according to the present embodiment, a storage device different from the storage unit 201 of the information processing device 103 may store the blood pressure estimation model database 220 . For example, when a server device (computer) including the storage device and the information processing device 103 are connected via a network, the information processing device 103 connects via the network to a blood pressure estimation model database stored in the storage device. From 220, a blood pressure estimation model 222 is obtained. For example, in this modified example, the network is the Internet, a LAN (Local Area Network), or the like.

(Modification 4)
As a modification 4 of the measurement system 100 according to the present embodiment, a server device (computer) different from the information processing device 103 may generate the blood pressure estimation model 222 by learning using machine learning or the like. The information processing device 103 may then use the blood pressure estimation model 222 generated by the server device to estimate the estimated blood pressure 240 through the processing described above.

(Second embodiment)
A second embodiment will be described with reference to FIGS. In the drawings, the same or equivalent elements are denoted by the same reference numerals, and overlapping descriptions are omitted. Configurations and processes having substantially the same functions as those of the first embodiment will be referred to by common reference numerals, description thereof will be omitted, and differences from the first embodiment will be described.

FIG. 10 is a diagram showing an example of the configuration of the information processing device 103 according to this embodiment. The difference between the information processing apparatus 103 illustrated in FIG. 10 and the information processing apparatus 103 illustrated in FIG. 2 is that the information processing apparatus 103 illustrated in FIG.

An evaluation index determination unit 1001 determines an index 1011 from the first feature quantity 237 . An index 1011 indicates the feature of the first feature quantity 237 . The index 1011 includes at least one selected from the group consisting of a blood pressure index for blood pressure and a pulse rate index for pulse rate.

The category selection unit 208 according to this embodiment selects a category 238 from a plurality of categories classified according to criteria related to the index 1011 . The criteria indicated by the criteria information 221 according to this embodiment include criteria related to the indicator 1011 .

FIG. 11 is a flowchart showing an example of processing for selecting the classification 238 by the information processing apparatus 103 according to this embodiment. It is assumed that the processing from step S501 illustrated in FIG. 5 to step S601 illustrated in FIG.

In step S<b>1101 , the evaluation index determination unit 1001 determines an index 1011 from a plurality of feature quantities indicated by the feature quantity set 236 . The physiological condition information 234 may contain noise when measuring the physiological condition of the subject 101 . Therefore, the plurality of feature quantities indicated by the feature quantity set 236 may contain noise caused by the noise. Therefore, the evaluation index determination unit 1001 determines an index 1011 indicating the features of the plurality of feature quantities indicated by the feature quantity set 236 so as to suppress noise included in the plurality of feature quantities.

For example, let feature set 236 include feature x1, feature _{x2 ,} and feature _x3 . Then, it is assumed that the feature quantity x1 indicates the _time required for _one pulse wave, the feature quantity x2 indicates the rising angle of the pulse wave, and the feature quantity _x3 indicates the age. Here, the feature quantity x1 and the feature quantity x2 indicate the duration and rising angle of _one beat, _respectively , with respect to a plurality of beats of the pulse wave. In that case, for example, the evaluation index determination unit 1001 performs principal component analysis on the _first feature quantity 237 including the feature quantity x1, the feature quantity x2 _, and the feature quantity _x3 . Determine _an index 1011 that includes the calculated eigenvalues z1 _, z2, and _z3 .

In step S<b>1102 , the classification selection unit 208 selects criteria to be determined from the criteria information 221 . In step S1103, the classification selection unit 208 determines whether or not the index 1011 related to the determination target criterion selected in step S1102 among the indices 1011 determined in step S1101 satisfies the determination target criterion. .

For example, if the criterion to be determined is a criterion related to a blood pressure index and the index 1011 includes the blood pressure index, the classification selection unit 208 determines whether the blood pressure index included in the index 1011 satisfies the criterion to be determined. do. Further, for example, if the criterion to be determined is a criterion related to the pulse rate index and the index 1011 includes the pulse rate index, the classification selection unit 208 determines that the pulse rate index included in the index 1011 satisfies the criterion to be determined. Determine whether or not

When noise is included in the measurement result when measuring the physiological condition of the subject 101, the classification selection unit 208 appropriately determines whether the first feature amount 237 satisfies the criteria indicated by the criteria information 221. It may not be possible. However, the classification selection unit 208 determines whether the index 1011 satisfies the determination target criteria, thereby suppressing erroneous determination caused by noise included in the measurement result when measuring the physiological state.

In step S1104, the classification selection unit 208 determines whether all the criteria indicated by the criteria information 221 have been determined. If it has not been determined whether the index 1011 satisfies the determination target criteria for all the criteria indicated by the criteria information 221, the control unit 202 returns the process to step S1102. In other words, the classification selection unit 208 repeats the processing from step S1102 to step S1103 until all criteria indicated by the criteria information 221 are determined. On the other hand, if it is determined whether or not the indicator 1011 satisfies the criterion of the determination target for all criteria indicated by the criterion information 221, the control unit 202 shifts the process to step 1105. FIG.

In step S<b>1105 , the classification selection unit 208 determines the classification 238 to which the subject 101 belongs from the determination results for all criteria indicated by the criteria information 221 . As a result, the classification selection unit 208 uses the index 1011 indicating the characteristics of the information about the living body of the person to be measured 101 to select the classification to which the person to be measured 101 belongs from a plurality of classifications classified according to the criteria indicated by the reference information 221. Classification can be selected. Then, the control unit 202 shifts the processing to step S701 illustrated in FIG.

As described above, the information processing apparatus 103 according to the present embodiment can suppress noise included in the measurement result when measuring the physiological state and estimate the blood pressure using the blood pressure estimation model 222 corresponding to the appropriate classification 238 . Therefore, the information processing apparatus 103 according to this embodiment can further estimate the blood pressure of each subject 101 .

Each process executed in the above embodiments is not limited to the processing mode illustrated in each embodiment. The functional blocks described above may be implemented using either a logic circuit (hardware) formed in an integrated circuit or the like, or software using a CPU. Each process executed in the above embodiments may be executed by multiple computers. For example, the processing executed by each functional block of the control unit 202 of the information processing apparatus 103 may be partly executed by another computer, or all of the processing may be executed by a plurality of computers. may

The present disclosure is not limited to the above-described embodiments, and various modifications are possible within the scope indicated in the claims. Embodiments obtained by appropriately combining technical means disclosed in different embodiments is also included in the technical scope of the present invention. Furthermore, new technical features can be formed by combining the technical means disclosed in each embodiment.

100 measurement system, 101 subject, 102 imaging device, 103 information processing device, 201 storage unit, 202 control unit, 203 light source adjustment unit, 204 body surface area extraction unit, 205 pulse wave calculation unit, 206 biological information acquisition unit, 207 feature amount determination unit 208 classification selection unit 209 feature amount selection unit 210 blood pressure estimation unit 220 blood pressure estimation model database 221 reference information 222 blood pressure estimation model 231 image 232 pulse wave 233 biological information 234 physiology state information, 235 attribute information, 236 feature amount set, 237 first feature amount, 238 classification, 239 second feature amount, 240 estimated blood pressure, 1001 evaluation index determination unit, 1011 index

Claims (12)

a biometric information acquiring unit that acquires biometric information of a person to be measured;
a feature amount determination unit that determines a plurality of feature amounts from the biometric information;
a classification selection unit that selects a classification to which the person to be measured belongs from a plurality of classifications classified according to criteria related to a first feature quantity included in the plurality of feature quantities;
a blood pressure estimating unit that inputs a second feature amount included in the plurality of feature amounts to the blood pressure estimation model selected according to the classification and outputs an estimated blood pressure;
Information processing device. 2. The information processing apparatus according to claim 1, wherein said second feature amount is a feature amount selected from said plurality of feature amounts according to said classification. 3. The information processing apparatus according to claim 1, wherein said blood pressure estimation model includes a blood pressure estimation formula including a coefficient selected according to said classification and a variable to which said second feature quantity is input. 4. The information processing apparatus according to claim 1, wherein said classification selection section selects said classification based on a relationship between said first feature amount and said reference. the biological information includes physiological state information indicating the physiological state of the subject;
The information processing apparatus according to any one of claims 1 to 4, wherein the plurality of feature amounts include feature amounts determined from the physiological condition information. the biological information includes attribute information indicating attributes of the person to be measured;
6. The information processing apparatus according to claim 5, wherein said plurality of feature amounts include feature amounts determined from said attribute information. The plurality of feature amounts include feature amounts related to at least one selected from the group consisting of blood pressure and pulse rate,
7. The information processing apparatus according to claim 1, wherein said reference includes at least one selected from the group consisting of blood pressure reference and pulse rate reference. 8. The information processing apparatus according to claim 7, wherein said first feature quantity includes a feature quantity related to at least one selected from the group consisting of said blood pressure reference and said pulse rate reference. further comprising an evaluation index determination unit that determines an index indicating the characteristics of the first feature quantity from the first feature quantity,
The information processing apparatus according to any one of claims 1 to 8, wherein said criteria include criteria related to said indicators. 10. The information processing apparatus according to claim 9, wherein said index includes at least one selected from the group consisting of a blood pressure index relating to blood pressure and a pulse rate index relating to pulse rate. a step of acquiring biological information of a person to be measured;
a step of determining a plurality of feature quantities from the biometric information;
selecting a category to which the person to be measured belongs from a plurality of categories classified according to a criterion related to a first feature included in the plurality of features;
a step of inputting a second feature amount included in the plurality of feature amounts into the blood pressure estimation model selected according to the classification and outputting an estimated blood pressure;
Blood pressure estimation method including. to the computer,
a function of acquiring biological information of a person to be measured;
a function of determining a plurality of feature amounts from the biometric information;
a function of selecting a category to which the person to be measured belongs from a plurality of categories classified according to a criterion related to a first feature included in the plurality of features;
a function of inputting a second feature quantity included in the plurality of feature quantities to a blood pressure estimation model selected according to the classification and outputting an estimated blood pressure;
program to run.

Images